羊群效应通常来说，是由于个别股票的暴涨或暴跌引起其他相关股票收 益率联动，致某类股票暴涨或暴跌。因此，通俗来说，羊群效应刻画的 是个股之间联动性的变化，并于此对趋势强度进行判断。
CCK模型由Chang, Cheng & Khorana 在 2000 年提出，其核心思想是通 过组*分股收益率相对于市场收益率Rm离散程度的变化识别羊群效 应的发生。
Chang, Cheng & Khorana 首先通过以下推导证明了理性情况下，由于个 股对市场风险的敏感程度不同，市场收益率Rm剧烈波动时，组合收益 率相对于Rm的离散程度会线性增加;
定义股票组合在t时刻的截面绝对离散度$CSAD_t$为
$$CSAD_t=\frac{1}{N}\sum_{i=1}^N|R_{i,t}-R_{m,t}|$$根据CAPM，有：
$$E(R_{i,t})=\gamma_0+\beta_iE(R_{m,t}-\gamma_0)$$则期望$CSAD_t$,即$E(CSAD_T)$为
$$E(CASD_t)=E(\frac{1}{N}\sum_{i=1}^N|R_{i,t}-R_{m,t}|)$$$=\frac{1}{N}(\sum_{i=1}^N|E(R_{i,t})-E(R_{m,t})|)$$=\frac{1}{N}(\sum_{i=1}^N|\gamma_0+\beta_iE(R_{i,t})-(\gamma_0+\beta_mE(R_{m,t}))|)$$=\frac{1}{N}\sum_{i=1}^N|\beta_i-\beta_m|E(R_{m,t}-\gamma)$对$E(CSAD_t)$求一，二阶导数，有
$$\frac{\delta E(CSAD_t)}{\delta E(R_{m,t})}=\frac{1}{N}\sum_{i=1}^N|\beta_i-\beta_m|>0$$$$\frac{\delta^2 E(CSAD_t)}{\delta E(R_{m,t})^2=0}$$一阶导数为正，二阶导数为零，说明理性情况，即无羊群效应时，$CSAD_t$和$R_m$的关系为线性正相关。
而存在羊群效应时，$CSAD_t$和$R_m$的线性正相关系数会被大破。基于这一思想，Chang,Cheng&Khorana构造了如下回归，根据回归中$R_{m,t}^2$的系数$\beta_2$是否显著为负判断是否存在羊群效应:$\beta_2$显著时，说明$CSAD_t$和$R_m$的关系为非线性；$\beta_2$为负时，随$R_m$增大，离散程度会减速上升或加速下降。减速上升说明$CSAD_t$上升幅度低于理想情况(理性情况匀速上升)，离散度加速下降更表明$CSAD_t$和$R_m$间存在强负相关关系。因此，$R_{m,t}^2$的系数$\beta_2$显著为负时说明羊群效应发生。
$CSAD_t=\alpha+\beta_1|R_{m,t}|+\beta_2R_{m,t}^2+\epsilon_t$CCK模型的所有因变量均基于市场收益率$R_m$构建，也即模型将市场作 为羊群效应的唯一驱动因素，但在真实市场中，风格、行业等多种因素 均会引起羊群效应，因此，模型变量可能需要根据市场实际情况进行调 整。接下来，我们将重点比较由不同驱动因素驱动的羊群所引起趋势的 相对强弱，并据此决定策略的具体调整方式。
从驱动因素的重要性出发，本文将主要探究市场驱动的羊群效应、市值 风格驱动的羊群效应发生后的市场：我们以 22 交易日（一个月，包括 当日）为滚动期，每天计算向前 22 交易日其上证 50 成分股组合截面绝 对离散度CSAD，估计以下两个模型的参数：
$CSAD_t=\alpha+\beta_1|R_{m,t}|+\beta_2R_{m,t}^2+\epsilon_t$$CSAD_t=\alpha+\beta_1|R_{smb,t}|+\beta_2 R_{smb,t}^2+\epsilon_t$其中，$R_{m,t}$指数收益率，$R_{smb,t}$为市值因子收益率，当对应二次项系数显著为负时，则发生了由该因素驱动的羊群效应。我们以滚动期内指 数平均日收益率的正负区分市场趋势为上涨还是下跌，上涨状态下各指数收盘价与羊群效应发生时间如下图所示。从图中可以看出：
羊群效应产生于趋势，作用于趋势，信号发出前后市场趋势明显。
多数趋势波段由市场、市值风格共同驱动，少数波段为单一因素驱动；
市场与市值风格多个“驱动力”共同驱动的羊群效应发生后，趋势更强：2018-2019 年间的羊群效应由市场或市值风格单个因素驱动，期间指数趋势较弱，甚至出现反转，其他年份则多由市场与市值风格共同作用，期间指数趋势更强。
import pandas as pdimport numpy as npimport statsmodels.api as smimport scipy.stats as stfrom jqdata import *from jqfactor import *import datetime as dtimport pickleimport matplotlib.pyplot as pltimport matplotlib.dates as mdatefrom IPython.core.display import HTML# 设置字体 用来正常显示中文标签mpl.rcParams['font.sans-serif'] = ['SimHei']# 用来正常显示负号plt.rcParams['axes.unicode_minus'] = False# 图表主题plt.style.use('ggplot')# 忽略报错import warningswarnings.filterwarnings("ignore")start = '2014-01-01'end = '2019-08-31'interval = 22
# 1 获取指数成分股# 过滤ST；过滤上市不足三个月=>filter_now_share；过滤当日停牌股票=>filter_paused_stocksdef filter_index_stocks(index_code, trade_date):# 获取成分股列表stocks = get_index_stocks(index_code, date=trade_date)# 剔除ST股st_stocks = get_extras('is_st', stocks, count=1, end_date=trade_date)stockList = [stock for stock in st_stocks if not st_stocks[stock][0]]# 剔除上市不足三月股票stockList = filter_now_share(stockList, trade_date)# 剔除当日停牌股票stockList = filter_paused_stocks(stockList, trade_date)return stockList# 2 过滤上市不足3月的股票def filter_now_share(stocks, begin_date, n=3 * 30):stockList = []# 如果begin_dateif type(begin_date) == str:begin_date = dt.datetime.strptime(begin_date, "%Y-%m-%d").date()for stock in stocks:start_date = get_security_info(stock).start_dateif start_date < (begin_date - dt.timedelta(days=n)):stockList.append(stock)return stockList# 3 过滤当日停牌股票def filter_paused_stocks(stockList, begin_date):is_paused = get_price(stockList, end_date=begin_date, count=1, fields='paused')['paused'].Tunsuspened_stocks = is_paused[is_paused.iloc[:, 0] < 1].index.tolist()return unsuspened_stocks# 4 获取基础数据'''index_code：为成分股代码start,end：str 日期interval为N日收益率file_name:储存文件的文件名实际运行时数据日期会在start向前推interval日return dict key['close','smb'] value:df index为date,columns为股票代码'''def get_datas(index_code, start, end, interval, file_name):# 向前推N日begin = get_trade_days(end_date=start, count=interval)[0]# 获取交易日利trade_list = get_trade_days(start_date=begin, end_date=end)datas = {}stock_df = []smb_list = []for date in trade_list:# 获取股票列表stocksList = filter_index_stocks(index_code, date)# 获取close数据stocks = get_price(stocksList, end_date=date, count=1, fields='close')['close']stock_df.append(stocks)# 获取smb数据q = query(valuation.code, valuation.day, valuation.circulating_market_cap).filter( valuation.code.in_(stocksList))smb = get_fundamentals(q, date=date)# 单位亿元smb = smb.pivot(index='day',columns='code',values='circulating_market_cap')smb_list.append(smb)print('success', date.strftime('%Y-%m-%d'))# 合并数据close_df = pd.concat(stock_df)smb_df = pd.concat(smb_list)datas['close'] = close_dfdatas['smb'] = smb_df# 存储数据pkl_file = open(file_name, 'wb')pickle.dump(datas, pkl_file)print('以储存数据：' + file_name)return datas# 5 获取信号'''datas_df：成分股数据，index为日期，columns为代码interval为间隔N日收益率,22为月收益率file_name:储存文件的文件名实际运行时数据日期会在start向前推interval日return df'''def get_factor(dic, index_code, interval):stocks = dic['close']smb_df = dic['smb']# 获取日期begin = min(hs300_datas['close'].index).strftime('%Y-%m-%d')end = max(hs300_datas['close'].index).strftime('%Y-%m-%d')index_close = get_price(index_code,start_date=begin,end_date=end,fields='close')['close']# 成分股# N-1实际为22日得收益率stock_ret = stocks.pct_change(interval - 1)stock_ret = stock_ret[interval - 1:]smb_ret = smb_df.pct_change(interval - 1)smb_ret = smb_ret[interval - 1:]# 指数index_Nret = index_close.pct_change(interval - 1)index_Nret = index_Nret[interval - 1:]# 指数日收益率index_ret = index_close.pct_change()index_ret = index_ret[interval - 1:]# 指数收盘价index_close = index_close[interval - 1:]# 将指数的一维数据转为与成分股相同的矩阵相减ret_diff_arr = stock_ret.values - np.broadcast_to(np.expand_dims(index_Nret.values, axis=1), (stock_ret.values.shape))smb_diff_arr = smb_ret.values - np.broadcast_to(np.expand_dims(index_Nret.values, axis=1), (smb_ret.values.shape))# 计算CSADcsad_arr = np.nansum(abs(ret_diff_arr), axis=1) / np.count_nonzero(ret_diff_arr, axis=1)csad_smb = np.nansum(abs(smb_diff_arr), axis=1) / np.count_nonzero(smb_diff_arr, axis=1)# 收益绝对值ret_arr = abs(stock_ret.values)smb_arr = abs(smb_ret.values)# 收益方ret_2_arr = ret_arr**2smb_2_arr = smb_arr**2# 回归temp = []for i in range(len(stock_ret)):trade_date = stock_ret.index[i]X_a = np.nan_to_num(np.c_[ret_arr[i], ret_2_arr[i]])Y_a = np.broadcast_to(np.expand_dims(csad_arr[i], axis=0), (len(X_a), 1))beta = np.linalg.lstsq(X_a, Y_a)[0][1][0] # 获取回归系数，最小二乘法X_b = np.nan_to_num(np.c_[smb_arr[i], smb_2_arr[i]])Y_b = np.broadcast_to(np.expand_dims(csad_smb[i], axis=0), (len(X_b), 1))beta_smb = np.linalg.lstsq(X_b, Y_b)[0][1][0]r_score = cal_score(beta, index_Nret[i])smb_score = cal_score(beta_smb, index_Nret[i])temp.append([beta, beta_smb, index_Nret[i], index_ret[i], index_close[i],beta < 0 and index_Nret[i] > 0, beta < 0 and index_Nret[i] < 0,beta_smb < 0 and index_Nret[i] < 0, r_score, smb_score])# 列名column_name = ['r_csad', 'smb_csad', 'index_Nret', '当日涨幅', 'close', 'r_up_singal','r_down_singal', 'smb_singal', 'r_factor', 'smb_factor']# 构建dffactor_df = pd.DataFrame(temp, columns=column_name, index=stock_ret.index)# 存储数据#pkl_file=open(file_name,'wb')#pickle.dump(factor,pkl_file)#print('以储存数据：'+file_name)return factor_df# 5-1 将CSAD和指数收益*打分def cal_score(csad, ret):if csad < 0 and ret > 0:score = csad + retelse:score=abs(csad)+abs(ret)return score# 6 回测函数'''输入：df index为日期,singal_col为df中含信号得列名-总体逻辑是有信号则买入持有，无信号则平仓'''def back_test(df, singal_col,holding=None, prt=False):position = []if holding == None:for i in range(len(df)):factor_value = df[singal_col][i]if factor_value:position.append(1)else:position.append(0)else:count = holdingfor i in range(len(df)):factor_value = df[singal_col][i]if factor_value:position.append(1)count = 1else:if count < holding:count += 1position.append(1)else:position.append(0)df['position'] = positionif prt:position = np.array(position)print('满仓天数：', len(position[position == 1]))print('空仓天数：', len(position[position == 0]))# 计算收益率index_ret = df.close.pct_change().values #df['当日收益率']ret = [0]# 确定哪些是开仓位置for i in range(len(df) - 1):ret.append(index_ret[i + 1] * position[i])ret = np.array(ret)df['ret'] = retcum_ret = []# 计算净值for i in range(len(ret)):if i == 0:cum_ret.append(1 + ret[i])else:cum_ret.append(cum_ret[-1] * (1 + ret[i]))df['cum_ret'] = cum_retreturn df# 7 生成回测报告def summary(df):#输出各项指标cum_ret = df['cum_ret']ret = df['ret']# 计算年华收益率annual_ret = cum_ret[-1]**(240 / (len(ret) - 5)) - 1# 计算累计收益率cum_ret_rate = cum_ret[-1] - 1# 最大回撤max_nv = np.maximum.accumulate(cum_ret)mdd = -np.min(cum_ret / max_nv - 1)print('年化收益率: {:.2%}'.format(annual_ret))print('累计收益率: {:.2%}'.format(cum_ret_rate))print('最大回撤: {:.2%}'.format(mdd))print('夏普比率：{:.2}'.format(ret.mean() / ret.std() * np.sqrt(240)))#作图plt.figure(1, figsize=(20, 10))plt.title('净值曲线', fontsize=18)plt.plot(df.index, cum_ret)plt.plot(df.index, df['close'] / df['close'][0])plt.legend(['策略净值', '基准净值'], fontsize=15)plt.figure(2, figsize=(20, 10))plt.title('相对优势', fontsize=18)plt.plot(df.index, cum_ret - df['close'] / df['close'][0])plt.show()# 分组回测'''df 为分组后的信号数据group_col 为有分组的列名holding 为持有天数'''def group_back_test(df, group_col, holding=None):group_list = df[group_col].unique().tolist()group_list.sort()ret_dic = {} # 储存每组回测收益率cum_ret_dic = {} # 储存每组净值report = {} # 储存每组报告数据# 获取指数每日收益率index_ret = df.close.pct_change().values# 获取指数收盘价index_close = df.close.valuesindex_close = index_close[:-1]# 基准净值cum_ret_dic['基准净值'] = index_close / index_close[0]for group_num in group_list:ret = [] # 储存收益率cum_ret = [] # 储存净值# 标注#-if holding == None:position = np.zeros(len(df))mask = df[group_col] == group_numposition[mask] = 1else:count = holdingposition = []for i in range(len(df)):threshold = df[group_col][i]if threshold == group_num:position.append(1)count = 1else:if count < holding:count += 1position.append(1)else:position.append(0)## 计算收益率for i in range(len(index_ret) - 1):# 取滞后一期得收益ret.append(index_ret[i + 1] * position[i])ret = np.array(ret)winning_count=np.sum(np.where(ret>0,1,0))/np.count_nonzero(position) # 日胜率ret_dic[group_num] = ret# 计算净值for i in range(len(ret)):if i == 0:cum_ret.append(1 + ret[i])else:cum_ret.append(cum_ret[-1] * (1 + ret[i]))cum_ret_dic[group_num] = cum_ret#-关键指标计算# 计算年化收益率annual_ret = cum_ret[-1]**(240 / (len(ret) - 5)) - 1# 计算累计收益率cum_ret_rate = cum_ret[-1] - 1# 最大回撤max_nv = np.maximum.accumulate(cum_ret)mdd = -np.min(cum_ret / max_nv - 1)# 储存每组报告数据report[group_num] = {'满仓天数': np.count_nonzero(position),'空仓天数': len(position) - np.count_nonzero(position),'日胜率':'{:.2%}'.format(winning_count),'年化收益率': '{:.2%}'.format(annual_ret),'累计收益率': '{:.2%}'.format(cum_ret_rate),'最大回撤': '{:.2%}'.format(mdd),'夏普比率': '{:.2}'.format(ret.mean() / ret.std() * np.sqrt(240))}return cum_ret_dic, report#输入参数data为包含因子值得原始数据集，num_group为组数，factor为用于排名的因子名称def get_group(data, num_group=5, factor='r_singal'):ranks = data[factor].rank(ascending=False) #按降序排名，组号越大，越好label = ['G' + str(i) for i in range(1, num_group + 1)] #创建组号category = pd.cut(ranks, bins=num_group, labels=label)category.name = 'GROUP'new_data = data.join(category) #将排名合并入原始数据集中return new_data# 敏感性分析def threshold_analysis(df, threshold_col, holding=[5, 10, 15, 20, 25], params='sharpe'):ranks = df[threshold_col].rank(ascending=False)df['SCORE'] = ranksnum = 5g = pd.cut(ranks.values, 5)threshold = g.categories.left.tolist()[1:]temp = []for i in threshold:for j in holding:bt = threshold_test(df, 'SCORE', i, j)if params == 'ret':cum_ret = bt['cum_ret']temp.append('{:.2%}'.format(cum_ret[-1]**(240 / (len(cum_ret) - 5)) -1))else:ret = bt['ret']temp.append(ret.mean() / ret.std() * np.sqrt(240))temp = np.array(temp).reshape((len(threshold), len(holding)))columns_name=list(map(lambda x:'持有{}天'.format(x),holding))temp_df = pd.DataFrame(temp, index=threshold, columns=columns_name)return temp_df# 敏感性分析回测用def threshold_test(df, singal_col, threshold, holding):count = holdingposition = []for i in range(len(df)):factor_value = df[singal_col][i]if factor_value > threshold:position.append(1)count = 1else:if count < holding:count += 1position.append(1)else:position.append(0)df['position'] = position# 计算收益率index_ret = df.close.pct_change().valuesret = [0]# 确定哪些是开仓位置for i in range(len(df) - 1):ret.append(index_ret[i + 1] * position[i])ret = np.array(ret)df['ret'] = retcum_ret = []# 计算净值for i in range(len(ret)):if i == 0:cum_ret.append(1 + ret[i])else:cum_ret.append(cum_ret[-1] * (1 + ret[i]))df['cum_ret'] = cum_retreturn df# 获取数据hs300_df = get_datas('000300.XSHG', start, end, interval, 'hs300_datas.pkl')zz500_df = get_datas('000905.XSHG', start, end, interval, 'zz500_datas.pkl')cyb_df = get_datas('399006.XSHE', start, end, interval, 'cyb_datas.pkl')sz50_df = get_datas('000016.XSHG', start, end, interval, 'sz50_datas.pkl')szzs_df=get_datas('000001.XSHG', start, end, interval, 'szzs_datas.pkl')zbzs_df=get_datas('399101.XSHE', start, end, interval, 'zbzs_datas.pkl')success 2013-12-02 success 2013-12-03 success 2013-12-04 success 2013-12-05 success 2013-12-06 success 2013-12-09 success 2013-12-10 success 2013-12-11 success 2013-12-12 success 2013-12-13 success 2013-12-16 success 2013-12-17 success 2013-12-18 success 2013-12-19 success 2013-12-20 success 2013-12-23 success 2013-12-24 success 2013-12-25 success 2013-12-26 success 2013-12-27 success 2013-12-30 success 2013-12-31 success 2014-01-02 success 2014-01-03 success 2014-01-06 success 2014-01-07 success 2014-01-08 success 2014-01-09 success 2014-01-10 success 2014-01-13 success 2014-01-14 success 2014-01-15 success 2014-01-16 success 2014-01-17 success 2014-01-20 success 2014-01-21 success 2014-01-22 success 2014-01-23 success 2014-01-24 success 2014-01-27 success 2014-01-28 success 2014-01-29 success 2014-01-30 success 2014-02-07 success 2014-02-10 success 2014-02-11 success 2014-02-12 success 2014-02-13 success 2014-02-14 success 2014-02-17 success 2014-02-18 success 2014-02-19 success 2014-02-20 success 2014-02-21 success 2014-02-24 success 2014-02-25 success 2014-02-26 success 2014-02-27 success 2014-02-28 success 2014-03-03 success 2014-03-04 success 2014-03-05 success 2014-03-06 success 2014-03-07 success 2014-03-10 success 2014-03-11 success 2014-03-12 success 2014-03-13 success 2014-03-14 success 2014-03-17 success 2014-03-18 success 2014-03-19 success 2014-03-20 success 2014-03-21 success 2014-03-24 success 2014-03-25 success 2014-03-26 success 2014-03-27 success 2014-03-28 success 2014-03-31 success 2014-04-01 success 2014-04-02 success 2014-04-03 success 2014-04-04 success 2014-04-08 success 2014-04-09 success 2014-04-10 success 2014-04-11 success 2014-04-14 success 2014-04-15 success 2014-04-16 success 2014-04-17 success 2014-04-18 success 2014-04-21 success 2014-04-22 success 2014-04-23 success 2014-04-24 success 2014-04-25 success 2014-04-28 success 2014-04-29 success 2014-04-30 success 2014-05-05 success 2014-05-06 success 2014-05-07 success 2014-05-08 success 2014-05-09 success 2014-05-12 success 2014-05-13 success 2014-05-14 success 2014-05-15 success 2014-05-16 success 2014-05-19 success 2014-05-20 success 2014-05-21 success 2014-05-22 success 2014-05-23 success 2014-05-26 success 2014-05-27 success 2014-05-28 success 2014-05-29 success 2014-05-30 success 2014-06-03 success 2014-06-04 success 2014-06-05 success 2014-06-06 success 2014-06-09 success 2014-06-10 success 2014-06-11 success 2014-06-12 success 2014-06-13 success 2014-06-16 success 2014-06-17 success 2014-06-18 success 2014-06-19 success 2014-06-20 success 2014-06-23 success 2014-06-24 success 2014-06-25 success 2014-06-26 success 2014-06-27 success 2014-06-30 success 2014-07-01 success 2014-07-02 success 2014-07-03 success 2014-07-04 success 2014-07-07 success 2014-07-08 success 2014-07-09 success 2014-07-10 success 2014-07-11 success 2014-07-14 success 2014-07-15 success 2014-07-16 success 2014-07-17 success 2014-07-18 success 2014-07-21 success 2014-07-22 success 2014-07-23 success 2014-07-24 success 2014-07-25 success 2014-07-28 success 2014-07-29 success 2014-07-30 success 2014-07-31 success 2014-08-01 success 2014-08-04 success 2014-08-05 success 2014-08-06 success 2014-08-07 success 2014-08-08 success 2014-08-11 success 2014-08-12 success 2014-08-13 success 2014-08-14 success 2014-08-15 success 2014-08-18 success 2014-08-19 success 2014-08-20 success 2014-08-21 success 2014-08-22 success 2014-08-25 success 2014-08-26 success 2014-08-27 success 2014-08-28 success 2014-08-29 success 2014-09-01 success 2014-09-02 success 2014-09-03 success 2014-09-04 success 2014-09-05 success 2014-09-09 success 2014-09-10 success 2014-09-11 success 2014-09-12 success 2014-09-15 success 2014-09-16 success 2014-09-17 success 2014-09-18 success 2014-09-19 success 2014-09-22 success 2014-09-23 success 2014-09-24 success 2014-09-25 success 2014-09-26 success 2014-09-29 success 2014-09-30 success 2014-10-08 success 2014-10-09 success 2014-10-10 success 2014-10-13 success 2014-10-14 success 2014-10-15 success 2014-10-16 success 2014-10-17 success 2014-10-20 success 2014-10-21 success 2014-10-22 success 2014-10-23 success 2014-10-24 success 2014-10-27 success 2014-10-28 success 2014-10-29 success 2014-10-30 success 2014-10-31 success 2014-11-03 success 2014-11-04 success 2014-11-05 success 2014-11-06 success 2014-11-07 success 2014-11-10 success 2014-11-11 success 2014-11-12 success 2014-11-13 success 2014-11-14 success 2014-11-17 success 2014-11-18 success 2014-11-19 success 2014-11-20 success 2014-11-21 success 2014-11-24 success 2014-11-25 success 2014-11-26 success 2014-11-27 success 2014-11-28 success 2014-12-01 success 2014-12-02 success 2014-12-03 success 2014-12-04 success 2014-12-05 success 2014-12-08 success 2014-12-09 success 2014-12-10 success 2014-12-11 success 2014-12-12 success 2014-12-15 success 2014-12-16 success 2014-12-17 success 2014-12-18 success 2014-12-19 success 2014-12-22 success 2014-12-23 success 2014-12-24 success 2014-12-25 success 2014-12-26 success 2014-12-29 success 2014-12-30 success 2014-12-31 success 2015-01-05 success 2015-01-06 success 2015-01-07 success 2015-01-08 success 2015-01-09 success 2015-01-12 success 2015-01-13 success 2015-01-14 success 2015-01-15 success 2015-01-16 success 2015-01-19 success 2015-01-20 success 2015-01-21 success 2015-01-22 success 2015-01-23 success 2015-01-26 success 2015-01-27 success 2015-01-28 success 2015-01-29 success 2015-01-30 success 2015-02-02 success 2015-02-03 success 2015-02-04 success 2015-02-05 success 2015-02-06 success 2015-02-09 success 2015-02-10 success 2015-02-11 success 2015-02-12 success 2015-02-13 success 2015-02-16 success 2015-02-17 success 2015-02-25 success 2015-02-26 success 2015-02-27 success 2015-03-02 success 2015-03-03 success 2015-03-04 success 2015-03-05 success 2015-03-06 success 2015-03-09 success 2015-03-10 success 2015-03-11 success 2015-03-12 success 2015-03-13 success 2015-03-16 success 2015-03-17 success 2015-03-18 success 2015-03-19 success 2015-03-20 success 2015-03-23 success 2015-03-24 success 2015-03-25 success 2015-03-26 success 2015-03-27 success 2015-03-30 success 2015-03-31 success 2015-04-01 success 2015-04-02 success 2015-04-03 success 2015-04-07 success 2015-04-08 success 2015-04-09 success 2015-04-10 success 2015-04-13 success 2015-04-14 success 2015-04-15 success 2015-04-16 success 2015-04-17 success 2015-04-20 success 2015-04-21 success 2015-04-22 success 2015-04-23 success 2015-04-24 success 2015-04-27 success 2015-04-28 success 2015-04-29 success 2015-04-30 success 2015-05-04 success 2015-05-05 success 2015-05-06 success 2015-05-07 success 2015-05-08 success 2015-05-11 success 2015-05-12 success 2015-05-13 success 2015-05-14 success 2015-05-15 success 2015-05-18 success 2015-05-19 success 2015-05-20 success 2015-05-21 success 2015-05-22 success 2015-05-25 success 2015-05-26 success 2015-05-27 success 2015-05-28 success 2015-05-29 success 2015-06-01 success 2015-06-02 success 2015-06-03 success 2015-06-04 success 2015-06-05 success 2015-06-08 success 2015-06-09 success 2015-06-10 success 2015-06-11 success 2015-06-12 success 2015-06-15 success 2015-06-16 success 2015-06-17 success 2015-06-18 success 2015-06-19 success 2015-06-23 success 2015-06-24 success 2015-06-25 success 2015-06-26 success 2015-06-29 success 2015-06-30 success 2015-07-01 success 2015-07-02 success 2015-07-03 success 2015-07-06 success 2015-07-07 success 2015-07-08 success 2015-07-09 success 2015-07-10 success 2015-07-13 success 2015-07-14 success 2015-07-15 success 2015-07-16 success 2015-07-17 success 2015-07-20 success 2015-07-21 success 2015-07-22 success 2015-07-23 success 2015-07-24 success 2015-07-27 success 2015-07-28 success 2015-07-29 success 2015-07-30 success 2015-07-31 success 2015-08-03 success 2015-08-04 success 2015-08-05 success 2015-08-06 success 2015-08-07 success 2015-08-10 success 2015-08-11 success 2015-08-12 success 2015-08-13 success 2015-08-14 success 2015-08-17 success 2015-08-18 success 2015-08-19 success 2015-08-20 success 2015-08-21 success 2015-08-24 success 2015-08-25 success 2015-08-26 success 2015-08-27 success 2015-08-28 success 2015-08-31 success 2015-09-01 success 2015-09-02 success 2015-09-07 success 2015-09-08 success 2015-09-09 success 2015-09-10 success 2015-09-11 success 2015-09-14 success 2015-09-15 success 2015-09-16 success 2015-09-17 success 2015-09-18 success 2015-09-21 success 2015-09-22 success 2015-09-23 success 2015-09-24 success 2015-09-25 success 2015-09-28 success 2015-09-29 success 2015-09-30 success 2015-10-08 success 2015-10-09 success 2015-10-12 success 2015-10-13 success 2015-10-14 success 2015-10-15 success 2015-10-16 success 2015-10-19 success 2015-10-20 success 2015-10-21 success 2015-10-22 success 2015-10-23 success 2015-10-26 success 2015-10-27 success 2015-10-28 success 2015-10-29 success 2015-10-30 success 2015-11-02 success 2015-11-03 success 2015-11-04 success 2015-11-05 success 2015-11-06 success 2015-11-09 success 2015-11-10 success 2015-11-11 success 2015-11-12 success 2015-11-13 success 2015-11-16 success 2015-11-17 success 2015-11-18 success 2015-11-19 success 2015-11-20 success 2015-11-23 success 2015-11-24 success 2015-11-25 success 2015-11-26 success 2015-11-27 success 2015-11-30 success 2015-12-01 success 2015-12-02 success 2015-12-03 success 2015-12-04 success 2015-12-07 success 2015-12-08 success 2015-12-09 success 2015-12-10 success 2015-12-11 success 2015-12-14 success 2015-12-15 success 2015-12-16 success 2015-12-17 success 2015-12-18 success 2015-12-21 success 2015-12-22 success 2015-12-23 success 2015-12-24 success 2015-12-25 success 2015-12-28 success 2015-12-29 success 2015-12-30 success 2015-12-31 success 2016-01-04 success 2016-01-05 success 2016-01-06 success 2016-01-07 success 2016-01-08 success 2016-01-11 success 2016-01-12 success 2016-01-13 success 2016-01-14 success 2016-01-15 success 2016-01-18 success 2016-01-19 success 2016-01-20 success 2016-01-21 success 2016-01-22 success 2016-01-25 success 2016-01-26 success 2016-01-27 success 2016-01-28 success 2016-01-29 success 2016-02-01 success 2016-02-02 success 2016-02-03 success 2016-02-04 success 2016-02-05 success 2016-02-15 success 2016-02-16 success 2016-02-17 success 2016-02-18 success 2016-02-19 success 2016-02-22 success 2016-02-23 success 2016-02-24 success 2016-02-25 success 2016-02-26 success 2016-02-29 success 2016-03-01 success 2016-03-02 success 2016-03-03 success 2016-03-04 success 2016-03-07 success 2016-03-08 success 2016-03-09 success 2016-03-10 success 2016-03-11 success 2016-03-14 success 2016-03-15 success 2016-03-16 success 2016-03-17 success 2016-03-18 success 2016-03-21 success 2016-03-22 success 2016-03-23 success 2016-03-24 success 2016-03-25 success 2016-03-28 success 2016-03-29 success 2016-03-30 success 2016-03-31 success 2016-04-01 success 2016-04-05 success 2016-04-06 success 2016-04-07 success 2016-04-08 success 2016-04-11 success 2016-04-12 success 2016-04-13 success 2016-04-14 success 2016-04-15 success 2016-04-18 success 2016-04-19 success 2016-04-20 success 2016-04-21 success 2016-04-22 success 2016-04-25 success 2016-04-26 success 2016-04-27 success 2016-04-28 success 2016-04-29 success 2016-05-03 success 2016-05-04 success 2016-05-05 success 2016-05-06 success 2016-05-09 success 2016-05-10 success 2016-05-11 success 2016-05-12 success 2016-05-13 success 2016-05-16 success 2016-05-17 success 2016-05-18 success 2016-05-19 success 2016-05-20 success 2016-05-23 success 2016-05-24 success 2016-05-25 success 2016-05-26 success 2016-05-27 success 2016-05-30 success 2016-05-31 success 2016-06-01 success 2016-06-02 success 2016-06-03 success 2016-06-06 success 2016-06-07 success 2016-06-08 success 2016-06-13 success 2016-06-14 success 2016-06-15 success 2016-06-16 success 2016-06-17 success 2016-06-20 success 2016-06-21 success 2016-06-22 success 2016-06-23 success 2016-06-24 success 2016-06-27 success 2016-06-28 success 2016-06-29 success 2016-06-30 success 2016-07-01 success 2016-07-04 success 2016-07-05 success 2016-07-06 success 2016-07-07 success 2016-07-08 success 2016-07-11 success 2016-07-12 success 2016-07-13 success 2016-07-14 success 2016-07-15 success 2016-07-18 success 2016-07-19 success 2016-07-20 success 2016-07-21 success 2016-07-22 success 2016-07-25 success 2016-07-26 success 2016-07-27 success 2016-07-28 success 2016-07-29 success 2016-08-01 success 2016-08-02 success 2016-08-03 success 2016-08-04 success 2016-08-05 success 2016-08-08 success 2016-08-09 success 2016-08-10 success 2016-08-11 success 2016-08-12 success 2016-08-15 success 2016-08-16 success 2016-08-17 success 2016-08-18 success 2016-08-19 success 2016-08-22 success 2016-08-23 success 2016-08-24 success 2016-08-25 success 2016-08-26 success 2016-08-29 success 2016-08-30 success 2016-08-31 success 2016-09-01 success 2016-09-02 success 2016-09-05 success 2016-09-06 success 2016-09-07 success 2016-09-08 success 2016-09-09 success 2016-09-12 success 2016-09-13 success 2016-09-14 success 2016-09-19 success 2016-09-20 success 2016-09-21 success 2016-09-22 success 2016-09-23 success 2016-09-26 success 2016-09-27 success 2016-09-28 success 2016-09-29 success 2016-09-30 success 2016-10-10 success 2016-10-11 success 2016-10-12 success 2016-10-13 success 2016-10-14 success 2016-10-17 success 2016-10-18 success 2016-10-19 success 2016-10-20 success 2016-10-21 success 2016-10-24 success 2016-10-25 success 2016-10-26 success 2016-10-27 success 2016-10-28 success 2016-10-31 success 2016-11-01 success 2016-11-02 success 2016-11-03 success 2016-11-04 success 2016-11-07 success 2016-11-08 success 2016-11-09 success 2016-11-10 success 2016-11-11 success 2016-11-14 success 2016-11-15 success 2016-11-16 success 2016-11-17 success 2016-11-18 success 2016-11-21 success 2016-11-22 success 2016-11-23 success 2016-11-24 success 2016-11-25 success 2016-11-28 success 2016-11-29 success 2016-11-30 success 2016-12-01 success 2016-12-02 success 2016-12-05 success 2016-12-06 success 2016-12-07 success 2016-12-08 success 2016-12-09 success 2016-12-12 success 2016-12-13 success 2016-12-14 success 2016-12-15 success 2016-12-16 success 2016-12-19 success 2016-12-20 success 2016-12-21 success 2016-12-22 success 2016-12-23 success 2016-12-26 success 2016-12-27 success 2016-12-28 success 2016-12-29 success 2016-12-30 success 2017-01-03 success 2017-01-04 success 2017-01-05 success 2017-01-06 success 2017-01-09 success 2017-01-10 success 2017-01-11 success 2017-01-12 success 2017-01-13 success 2017-01-16 success 2017-01-17 success 2017-01-18 success 2017-01-19 success 2017-01-20 success 2017-01-23 success 2017-01-24 success 2017-01-25 success 2017-01-26 success 2017-02-03 success 2017-02-06 success 2017-02-07 success 2017-02-08 success 2017-02-09 success 2017-02-10 success 2017-02-13 success 2017-02-14 success 2017-02-15 success 2017-02-16 success 2017-02-17 success 2017-02-20 success 2017-02-21 success 2017-02-22 success 2017-02-23 success 2017-02-24 success 2017-02-27 success 2017-02-28 success 2017-03-01 success 2017-03-02 success 2017-03-03 success 2017-03-06 success 2017-03-07 success 2017-03-08 success 2017-03-09 success 2017-03-10 success 2017-03-13 success 2017-03-14 success 2017-03-15 success 2017-03-16 success 2017-03-17 success 2017-03-20 success 2017-03-21 success 2017-03-22 success 2017-03-23 success 2017-03-24 success 2017-03-27 success 2017-03-28 success 2017-03-29 success 2017-03-30 success 2017-03-31 success 2017-04-05 success 2017-04-06 success 2017-04-07 success 2017-04-10 success 2017-04-11 success 2017-04-12 success 2017-04-13 success 2017-04-14 success 2017-04-17 success 2017-04-18 success 2017-04-19 success 2017-04-20 success 2017-04-21 success 2017-04-24 success 2017-04-25 success 2017-04-26 success 2017-04-27 success 2017-04-28 success 2017-05-02 success 2017-05-03 success 2017-05-04 success 2017-05-05 success 2017-05-08 success 2017-05-09 success 2017-05-10 success 2017-05-11 success 2017-05-12 success 2017-05-15 success 2017-05-16 success 2017-05-17 success 2017-05-18 success 2017-05-19 success 2017-05-22 success 2017-05-23 success 2017-05-24 success 2017-05-25 success 2017-05-26 success 2017-05-31 success 2017-06-01 success 2017-06-02 success 2017-06-05 success 2017-06-06 success 2017-06-07 success 2017-06-08 success 2017-06-09 success 2017-06-12 success 2017-06-13 success 2017-06-14 success 2017-06-15 success 2017-06-16 success 2017-06-19 success 2017-06-20 success 2017-06-21 success 2017-06-22 success 2017-06-23 success 2017-06-26 success 2017-06-27 success 2017-06-28 success 2017-06-29 success 2017-06-30 success 2017-07-03 success 2017-07-04 success 2017-07-05 success 2017-07-06 success 2017-07-07 success 2017-07-10 success 2017-07-11 success 2017-07-12 success 2017-07-13 success 2017-07-14 success 2017-07-17 success 2017-07-18 success 2017-07-19 success 2017-07-20 success 2017-07-21 success 2017-07-24 success 2017-07-25 success 2017-07-26 success 2017-07-27 success 2017-07-28 success 2017-07-31 success 2017-08-01 success 2017-08-02 success 2017-08-03 success 2017-08-04 success 2017-08-07 success 2017-08-08 success 2017-08-09 success 2017-08-10 success 2017-08-11 success 2017-08-14 success 2017-08-15 success 2017-08-16 success 2017-08-17 success 2017-08-18 success 2017-08-21 success 2017-08-22 success 2017-08-23 success 2017-08-24 success 2017-08-25 success 2017-08-28 success 2017-08-29 success 2017-08-30 success 2017-08-31 success 2017-09-01 success 2017-09-04 success 2017-09-05 success 2017-09-06 success 2017-09-07 success 2017-09-08 success 2017-09-11 success 2017-09-12 success 2017-09-13 success 2017-09-14 success 2017-09-15 success 2017-09-18 success 2017-09-19 success 2017-09-20 success 2017-09-21 success 2017-09-22 success 2017-09-25 success 2017-09-26 success 2017-09-27 success 2017-09-28 success 2017-09-29 success 2017-10-09 success 2017-10-10 success 2017-10-11 success 2017-10-12 success 2017-10-13 success 2017-10-16 success 2017-10-17 success 2017-10-18 success 2017-10-19 success 2017-10-20 success 2017-10-23 success 2017-10-24 success 2017-10-25 success 2017-10-26 success 2017-10-27 success 2017-10-30 success 2017-10-31 success 2017-11-01 success 2017-11-02 success 2017-11-03 success 2017-11-06 success 2017-11-07 success 2017-11-08 success 2017-11-09 success 2017-11-10 success 2017-11-13 success 2017-11-14 success 2017-11-15 success 2017-11-16 success 2017-11-17 success 2017-11-20 success 2017-11-21 success 2017-11-22 success 2017-11-23 success 2017-11-24 success 2017-11-27 success 2017-11-28 success 2017-11-29 success 2017-11-30 success 2017-12-01 success 2017-12-04 success 2017-12-05 success 2017-12-06 success 2017-12-07 success 2017-12-08 success 2017-12-11 success 2017-12-12 success 2017-12-13 success 2017-12-14 success 2017-12-15 success 2017-12-18 success 2017-12-19 success 2017-12-20 success 2017-12-21 success 2017-12-22 success 2017-12-25 success 2017-12-26 success 2017-12-27 success 2017-12-28 success 2017-12-29 success 2018-01-02 success 2018-01-03 success 2018-01-04 success 2018-01-05 success 2018-01-08 success 2018-01-09 success 2018-01-10 success 2018-01-11 success 2018-01-12 success 2018-01-15 success 2018-01-16 success 2018-01-17 success 2018-01-18 success 2018-01-19 success 2018-01-22 success 2018-01-23 success 2018-01-24 success 2018-01-25 success 2018-01-26 success 2018-01-29 success 2018-01-30 success 2018-01-31 success 2018-02-01 success 2018-02-02 success 2018-02-05 success 2018-02-06 success 2018-02-07 success 2018-02-08 success 2018-02-09 success 2018-02-12 success 2018-02-13 success 2018-02-14 success 2018-02-22 success 2018-02-23 success 2018-02-26 success 2018-02-27 success 2018-02-28 success 2018-03-01 success 2018-03-02 success 2018-03-05 success 2018-03-06 success 2018-03-07 success 2018-03-08 success 2018-03-09 success 2018-03-12 success 2018-03-13 success 2018-03-14 success 2018-03-15 success 2018-03-16 success 2018-03-19 success 2018-03-20 success 2018-03-21 success 2018-03-22 success 2018-03-23 success 2018-03-26 success 2018-03-27 success 2018-03-28 success 2018-03-29 success 2018-03-30 success 2018-04-02 success 2018-04-03 success 2018-04-04 success 2018-04-09 success 2018-04-10 success 2018-04-11 success 2018-04-12 success 2018-04-13 success 2018-04-16 success 2018-04-17 success 2018-04-18 success 2018-04-19 success 2018-04-20 success 2018-04-23 success 2018-04-24 success 2018-04-25 success 2018-04-26 success 2018-04-27 success 2018-05-02 success 2018-05-03 success 2018-05-04 success 2018-05-07 success 2018-05-08 success 2018-05-09 success 2018-05-10 success 2018-05-11 success 2018-05-14 success 2018-05-15 success 2018-05-16 success 2018-05-17 success 2018-05-18 success 2018-05-21 success 2018-05-22 success 2018-05-23 success 2018-05-24 success 2018-05-25 success 2018-05-28 success 2018-05-29 success 2018-05-30 success 2018-05-31 success 2018-06-01 success 2018-06-04 success 2018-06-05 success 2018-06-06 success 2018-06-07 success 2018-06-08 success 2018-06-11 success 2018-06-12 success 2018-06-13 success 2018-06-14 success 2018-06-15 success 2018-06-19 success 2018-06-20 success 2018-06-21 success 2018-06-22 success 2018-06-25 success 2018-06-26 success 2018-06-27 success 2018-06-28 success 2018-06-29 success 2018-07-02 success 2018-07-03 success 2018-07-04 success 2018-07-05 success 2018-07-06 success 2018-07-09 success 2018-07-10 success 2018-07-11 success 2018-07-12 success 2018-07-13 success 2018-07-16 success 2018-07-17 success 2018-07-18 success 2018-07-19 success 2018-07-20 success 2018-07-23 success 2018-07-24 success 2018-07-25 success 2018-07-26 success 2018-07-27 success 2018-07-30 success 2018-07-31 success 2018-08-01 success 2018-08-02 success 2018-08-03 success 2018-08-06 success 2018-08-07 success 2018-08-08 success 2018-08-09 success 2018-08-10 success 2018-08-13 success 2018-08-14 success 2018-08-15 success 2018-08-16 success 2018-08-17 success 2018-08-20 success 2018-08-21 success 2018-08-22 success 2018-08-23 success 2018-08-24 success 2018-08-27 success 2018-08-28 success 2018-08-29 success 2018-08-30 success 2018-08-31 success 2018-09-03 success 2018-09-04 success 2018-09-05 success 2018-09-06 success 2018-09-07 success 2018-09-10 success 2018-09-11 success 2018-09-12 success 2018-09-13 success 2018-09-14 success 2018-09-17 success 2018-09-18 success 2018-09-19 success 2018-09-20 success 2018-09-21 success 2018-09-25 success 2018-09-26 success 2018-09-27 success 2018-09-28 success 2018-10-08 success 2018-10-09 success 2018-10-10 success 2018-10-11 success 2018-10-12 success 2018-10-15 success 2018-10-16 success 2018-10-17 success 2018-10-18 success 2018-10-19 success 2018-10-22 success 2018-10-23 success 2018-10-24 success 2018-10-25 success 2018-10-26 success 2018-10-29 success 2018-10-30 success 2018-10-31 success 2018-11-01 success 2018-11-02 success 2018-11-05 success 2018-11-06 success 2018-11-07 success 2018-11-08 success 2018-11-09 success 2018-11-12 success 2018-11-13 success 2018-11-14 success 2018-11-15 success 2018-11-16 success 2018-11-19 success 2018-11-20 success 2018-11-21 success 2018-11-22 success 2018-11-23 success 2018-11-26 success 2018-11-27 success 2018-11-28 success 2018-11-29 success 2018-11-30 success 2018-12-03 success 2018-12-04 success 2018-12-05 success 2018-12-06 success 2018-12-07 success 2018-12-10 success 2018-12-11 success 2018-12-12 success 2018-12-13 success 2018-12-14 success 2018-12-17 success 2018-12-18 success 2018-12-19 success 2018-12-20 success 2018-12-21 success 2018-12-24 success 2018-12-25 success 2018-12-26 success 2018-12-27 success 2018-12-28 success 2019-01-02 success 2019-01-03 success 2019-01-04 success 2019-01-07 success 2019-01-08 success 2019-01-09 success 2019-01-10 success 2019-01-11 success 2019-01-14 success 2019-01-15 success 2019-01-16 success 2019-01-17 success 2019-01-18 success 2019-01-21 success 2019-01-22 success 2019-01-23 success 2019-01-24 success 2019-01-25 success 2019-01-28 success 2019-01-29 success 2019-01-30 success 2019-01-31 success 2019-02-01 success 2019-02-11 success 2019-02-12 success 2019-02-13 success 2019-02-14 success 2019-02-15 success 2019-02-18 success 2019-02-19 success 2019-02-20 success 2019-02-21 success 2019-02-22 success 2019-02-25 success 2019-02-26 success 2019-02-27 success 2019-02-28 success 2019-03-01 success 2019-03-04 success 2019-03-05 success 2019-03-06 success 2019-03-07 success 2019-03-08 success 2019-03-11 success 2019-03-12 success 2019-03-13 success 2019-03-14 success 2019-03-15 success 2019-03-18 success 2019-03-19 success 2019-03-20 success 2019-03-21 success 2019-03-22 success 2019-03-25 success 2019-03-26 success 2019-03-27 success 2019-03-28 success 2019-03-29 success 2019-04-01 success 2019-04-02 success 2019-04-03 success 2019-04-04 success 2019-04-08 success 2019-04-09 success 2019-04-10 success 2019-04-11 success 2019-04-12 success 2019-04-15 success 2019-04-16 success 2019-04-17 success 2019-04-18 success 2019-04-19 success 2019-04-22 success 2019-04-23 success 2019-04-24 success 2019-04-25 success 2019-04-26 success 2019-04-29 success 2019-04-30 success 2019-05-06 success 2019-05-07 success 2019-05-08 success 2019-05-09 success 2019-05-10 success 2019-05-13 success 2019-05-14 success 2019-05-15 success 2019-05-16 success 2019-05-17 success 2019-05-20 success 2019-05-21 success 2019-05-22 success 2019-05-23 success 2019-05-24 success 2019-05-27 success 2019-05-28 success 2019-05-29 success 2019-05-30 success 2019-05-31 success 2019-06-03 success 2019-06-04 success 2019-06-05 success 2019-06-06 success 2019-06-10 success 2019-06-11 success 2019-06-12 success 2019-06-13 success 2019-06-14 success 2019-06-17 success 2019-06-18 success 2019-06-19 success 2019-06-20 success 2019-06-21 success 2019-06-24 success 2019-06-25 success 2019-06-26 success 2019-06-27 success 2019-06-28 success 2019-07-01 success 2019-07-02 success 2019-07-03 success 2019-07-04 success 2019-07-05 success 2019-07-08 success 2019-07-09 success 2019-07-10 success 2019-07-11 success 2019-07-12 success 2019-07-15 success 2019-07-16 success 2019-07-17 success 2019-07-18 success 2019-07-19 success 2019-07-22 success 2019-07-23 success 2019-07-24 success 2019-07-25 success 2019-07-26 success 2019-07-29 success 2019-07-30 success 2019-07-31 success 2019-08-01 success 2019-08-02 success 2019-08-05 success 2019-08-06 success 2019-08-07 success 2019-08-08 success 2019-08-09 success 2019-08-12 success 2019-08-13 success 2019-08-14 success 2019-08-15 success 2019-08-16 success 2019-08-19 success 2019-08-20 success 2019-08-21 success 2019-08-22 success 2019-08-23 success 2019-08-26 success 2019-08-27 success 2019-08-28 success 2019-08-29 success 2019-08-30 以储存数据：szzs_datas.pkl success 2013-12-02 success 2013-12-03 success 2013-12-04 success 2013-12-05 success 2013-12-06 success 2013-12-09 success 2013-12-10 success 2013-12-11 success 2013-12-12 success 2013-12-13 success 2013-12-16 success 2013-12-17 success 2013-12-18 success 2013-12-19 success 2013-12-20 success 2013-12-23 success 2013-12-24 success 2013-12-25 success 2013-12-26 success 2013-12-27 success 2013-12-30 success 2013-12-31 success 2014-01-02 success 2014-01-03 success 2014-01-06 success 2014-01-07 success 2014-01-08 success 2014-01-09 success 2014-01-10 success 2014-01-13 success 2014-01-14 success 2014-01-15 success 2014-01-16 success 2014-01-17 success 2014-01-20 success 2014-01-21 success 2014-01-22 success 2014-01-23 success 2014-01-24 success 2014-01-27 success 2014-01-28 success 2014-01-29 success 2014-01-30 success 2014-02-07 success 2014-02-10 success 2014-02-11 success 2014-02-12 success 2014-02-13 success 2014-02-14 success 2014-02-17 success 2014-02-18 success 2014-02-19 success 2014-02-20 success 2014-02-21 success 2014-02-24 success 2014-02-25 success 2014-02-26 success 2014-02-27 success 2014-02-28 success 2014-03-03 success 2014-03-04 success 2014-03-05 success 2014-03-06 success 2014-03-07 success 2014-03-10 success 2014-03-11 success 2014-03-12 success 2014-03-13 success 2014-03-14 success 2014-03-17 success 2014-03-18 success 2014-03-19 success 2014-03-20 success 2014-03-21 success 2014-03-24 success 2014-03-25 success 2014-03-26 success 2014-03-27 success 2014-03-28 success 2014-03-31 success 2014-04-01 success 2014-04-02 success 2014-04-03 success 2014-04-04 success 2014-04-08 success 2014-04-09 success 2014-04-10 success 2014-04-11 success 2014-04-14 success 2014-04-15 success 2014-04-16 success 2014-04-17 success 2014-04-18 success 2014-04-21 success 2014-04-22 success 2014-04-23 success 2014-04-24 success 2014-04-25 success 2014-04-28 success 2014-04-29 success 2014-04-30 success 2014-05-05 success 2014-05-06 success 2014-05-07 success 2014-05-08 success 2014-05-09 success 2014-05-12 success 2014-05-13 success 2014-05-14 success 2014-05-15 success 2014-05-16 success 2014-05-19 success 2014-05-20 success 2014-05-21 success 2014-05-22 success 2014-05-23 success 2014-05-26 success 2014-05-27 success 2014-05-28 success 2014-05-29 success 2014-05-30 success 2014-06-03 success 2014-06-04 success 2014-06-05 success 2014-06-06 success 2014-06-09 success 2014-06-10 success 2014-06-11 success 2014-06-12 success 2014-06-13 success 2014-06-16 success 2014-06-17 success 2014-06-18 success 2014-06-19 success 2014-06-20 success 2014-06-23 success 2014-06-24 success 2014-06-25 success 2014-06-26 success 2014-06-27 success 2014-06-30 success 2014-07-01 success 2014-07-02 success 2014-07-03 success 2014-07-04 success 2014-07-07 success 2014-07-08 success 2014-07-09 success 2014-07-10 success 2014-07-11 success 2014-07-14 success 2014-07-15 success 2014-07-16 success 2014-07-17 success 2014-07-18 success 2014-07-21 success 2014-07-22 success 2014-07-23 success 2014-07-24 success 2014-07-25 success 2014-07-28 success 2014-07-29 success 2014-07-30 success 2014-07-31 success 2014-08-01 success 2014-08-04 success 2014-08-05 success 2014-08-06 success 2014-08-07 success 2014-08-08 success 2014-08-11 success 2014-08-12 success 2014-08-13 success 2014-08-14 success 2014-08-15 success 2014-08-18 success 2014-08-19 success 2014-08-20 success 2014-08-21 success 2014-08-22 success 2014-08-25 success 2014-08-26 success 2014-08-27 success 2014-08-28 success 2014-08-29 success 2014-09-01 success 2014-09-02 success 2014-09-03 success 2014-09-04 success 2014-09-05 success 2014-09-09 success 2014-09-10 success 2014-09-11 success 2014-09-12 success 2014-09-15 success 2014-09-16 success 2014-09-17 success 2014-09-18 success 2014-09-19 success 2014-09-22 success 2014-09-23 success 2014-09-24 success 2014-09-25 success 2014-09-26 success 2014-09-29 success 2014-09-30 success 2014-10-08 success 2014-10-09 success 2014-10-10 success 2014-10-13 success 2014-10-14 success 2014-10-15 success 2014-10-16 success 2014-10-17 success 2014-10-20 success 2014-10-21 success 2014-10-22 success 2014-10-23 success 2014-10-24 success 2014-10-27 success 2014-10-28 success 2014-10-29 success 2014-10-30 success 2014-10-31 success 2014-11-03 success 2014-11-04 success 2014-11-05 success 2014-11-06 success 2014-11-07 success 2014-11-10 success 2014-11-11 success 2014-11-12 success 2014-11-13 success 2014-11-14 success 2014-11-17 success 2014-11-18 success 2014-11-19 success 2014-11-20 success 2014-11-21 success 2014-11-24 success 2014-11-25 success 2014-11-26 success 2014-11-27 success 2014-11-28 success 2014-12-01 success 2014-12-02 success 2014-12-03 success 2014-12-04 success 2014-12-05 success 2014-12-08 success 2014-12-09 success 2014-12-10 success 2014-12-11 success 2014-12-12 success 2014-12-15 success 2014-12-16 success 2014-12-17 success 2014-12-18 success 2014-12-19 success 2014-12-22 success 2014-12-23 success 2014-12-24 success 2014-12-25 success 2014-12-26 success 2014-12-29 success 2014-12-30 success 2014-12-31 success 2015-01-05 success 2015-01-06 success 2015-01-07 success 2015-01-08 success 2015-01-09 success 2015-01-12 success 2015-01-13 success 2015-01-14 success 2015-01-15 success 2015-01-16 success 2015-01-19 success 2015-01-20 success 2015-01-21 success 2015-01-22 success 2015-01-23 success 2015-01-26 success 2015-01-27 success 2015-01-28 success 2015-01-29 success 2015-01-30 success 2015-02-02 success 2015-02-03 success 2015-02-04 success 2015-02-05 success 2015-02-06 success 2015-02-09 success 2015-02-10 success 2015-02-11 success 2015-02-12 success 2015-02-13 success 2015-02-16 success 2015-02-17 success 2015-02-25 success 2015-02-26 success 2015-02-27 success 2015-03-02 success 2015-03-03 success 2015-03-04 success 2015-03-05 success 2015-03-06 success 2015-03-09 success 2015-03-10 success 2015-03-11 success 2015-03-12 success 2015-03-13 success 2015-03-16 success 2015-03-17 success 2015-03-18 success 2015-03-19 success 2015-03-20 success 2015-03-23 success 2015-03-24 success 2015-03-25 success 2015-03-26 success 2015-03-27 success 2015-03-30 success 2015-03-31 success 2015-04-01 success 2015-04-02 success 2015-04-03 success 2015-04-07 success 2015-04-08 success 2015-04-09 success 2015-04-10 success 2015-04-13 success 2015-04-14 success 2015-04-15 success 2015-04-16 success 2015-04-17 success 2015-04-20 success 2015-04-21 success 2015-04-22 success 2015-04-23 success 2015-04-24 success 2015-04-27 success 2015-04-28 success 2015-04-29 success 2015-04-30 success 2015-05-04 success 2015-05-05 success 2015-05-06 success 2015-05-07 success 2015-05-08 success 2015-05-11 success 2015-05-12 success 2015-05-13 success 2015-05-14 success 2015-05-15 success 2015-05-18 success 2015-05-19 success 2015-05-20 success 2015-05-21 success 2015-05-22 success 2015-05-25 success 2015-05-26 success 2015-05-27 success 2015-05-28 success 2015-05-29 success 2015-06-01 success 2015-06-02 success 2015-06-03 success 2015-06-04 success 2015-06-05 success 2015-06-08 success 2015-06-09 success 2015-06-10 success 2015-06-11 success 2015-06-12 success 2015-06-15 success 2015-06-16 success 2015-06-17 success 2015-06-18 success 2015-06-19 success 2015-06-23 success 2015-06-24 success 2015-06-25 success 2015-06-26 success 2015-06-29 success 2015-06-30 success 2015-07-01 success 2015-07-02 success 2015-07-03 success 2015-07-06 success 2015-07-07 success 2015-07-08 success 2015-07-09 success 2015-07-10 success 2015-07-13 success 2015-07-14 success 2015-07-15 success 2015-07-16 success 2015-07-17 success 2015-07-20 success 2015-07-21 success 2015-07-22 success 2015-07-23 success 2015-07-24 success 2015-07-27 success 2015-07-28 success 2015-07-29 success 2015-07-30 success 2015-07-31 success 2015-08-03 success 2015-08-04 success 2015-08-05 success 2015-08-06 success 2015-08-07 success 2015-08-10 success 2015-08-11 success 2015-08-12 success 2015-08-13 success 2015-08-14 success 2015-08-17 success 2015-08-18 success 2015-08-19 success 2015-08-20 success 2015-08-21 success 2015-08-24 success 2015-08-25 success 2015-08-26 success 2015-08-27 success 2015-08-28 success 2015-08-31 success 2015-09-01 success 2015-09-02 success 2015-09-07 success 2015-09-08 success 2015-09-09 success 2015-09-10 success 2015-09-11 success 2015-09-14 success 2015-09-15 success 2015-09-16 success 2015-09-17 success 2015-09-18 success 2015-09-21 success 2015-09-22 success 2015-09-23 success 2015-09-24 success 2015-09-25 success 2015-09-28 success 2015-09-29 success 2015-09-30 success 2015-10-08 success 2015-10-09 success 2015-10-12 success 2015-10-13 success 2015-10-14 success 2015-10-15 success 2015-10-16 success 2015-10-19 success 2015-10-20 success 2015-10-21 success 2015-10-22 success 2015-10-23 success 2015-10-26 success 2015-10-27 success 2015-10-28 success 2015-10-29 success 2015-10-30 success 2015-11-02 success 2015-11-03 success 2015-11-04 success 2015-11-05 success 2015-11-06 success 2015-11-09 success 2015-11-10 success 2015-11-11 success 2015-11-12 success 2015-11-13 success 2015-11-16 success 2015-11-17 success 2015-11-18 success 2015-11-19 success 2015-11-20 success 2015-11-23 success 2015-11-24 success 2015-11-25 success 2015-11-26 success 2015-11-27 success 2015-11-30 success 2015-12-01 success 2015-12-02 success 2015-12-03 success 2015-12-04 success 2015-12-07 success 2015-12-08 success 2015-12-09 success 2015-12-10 success 2015-12-11 success 2015-12-14 success 2015-12-15 success 2015-12-16 success 2015-12-17 success 2015-12-18 success 2015-12-21 success 2015-12-22 success 2015-12-23 success 2015-12-24 success 2015-12-25 success 2015-12-28 success 2015-12-29 success 2015-12-30 success 2015-12-31 success 2016-01-04 success 2016-01-05 success 2016-01-06 success 2016-01-07 success 2016-01-08 success 2016-01-11 success 2016-01-12 success 2016-01-13 success 2016-01-14 success 2016-01-15 success 2016-01-18 success 2016-01-19 success 2016-01-20 success 2016-01-21 success 2016-01-22 success 2016-01-25 success 2016-01-26 success 2016-01-27 success 2016-01-28 success 2016-01-29 success 2016-02-01 success 2016-02-02 success 2016-02-03 success 2016-02-04 success 2016-02-05 success 2016-02-15 success 2016-02-16 success 2016-02-17 success 2016-02-18 success 2016-02-19 success 2016-02-22 success 2016-02-23 success 2016-02-24 success 2016-02-25 success 2016-02-26 success 2016-02-29 success 2016-03-01 success 2016-03-02 success 2016-03-03 success 2016-03-04 success 2016-03-07 success 2016-03-08 success 2016-03-09 success 2016-03-10 success 2016-03-11 success 2016-03-14 success 2016-03-15 success 2016-03-16 success 2016-03-17 success 2016-03-18 success 2016-03-21 success 2016-03-22 success 2016-03-23 success 2016-03-24 success 2016-03-25 success 2016-03-28 success 2016-03-29 success 2016-03-30 success 2016-03-31 success 2016-04-01 success 2016-04-05 success 2016-04-06 success 2016-04-07 success 2016-04-08 success 2016-04-11 success 2016-04-12 success 2016-04-13 success 2016-04-14 success 2016-04-15 success 2016-04-18 success 2016-04-19 success 2016-04-20 success 2016-04-21 success 2016-04-22 success 2016-04-25 success 2016-04-26 success 2016-04-27 success 2016-04-28 success 2016-04-29 success 2016-05-03 success 2016-05-04 success 2016-05-05 success 2016-05-06 success 2016-05-09 success 2016-05-10 success 2016-05-11 success 2016-05-12 success 2016-05-13 success 2016-05-16 success 2016-05-17 success 2016-05-18 success 2016-05-19 success 2016-05-20 success 2016-05-23 success 2016-05-24 success 2016-05-25 success 2016-05-26 success 2016-05-27 success 2016-05-30 success 2016-05-31 success 2016-06-01 success 2016-06-02 success 2016-06-03 success 2016-06-06 success 2016-06-07 success 2016-06-08 success 2016-06-13 success 2016-06-14 success 2016-06-15 success 2016-06-16 success 2016-06-17 success 2016-06-20 success 2016-06-21 success 2016-06-22 success 2016-06-23 success 2016-06-24 success 2016-06-27 success 2016-06-28 success 2016-06-29 success 2016-06-30 success 2016-07-01 success 2016-07-04 success 2016-07-05 success 2016-07-06 success 2016-07-07 success 2016-07-08 success 2016-07-11 success 2016-07-12 success 2016-07-13 success 2016-07-14 success 2016-07-15 success 2016-07-18 success 2016-07-19 success 2016-07-20 success 2016-07-21 success 2016-07-22 success 2016-07-25 success 2016-07-26 success 2016-07-27 success 2016-07-28 success 2016-07-29 success 2016-08-01 success 2016-08-02 success 2016-08-03 success 2016-08-04 success 2016-08-05 success 2016-08-08 success 2016-08-09 success 2016-08-10 success 2016-08-11 success 2016-08-12 success 2016-08-15 success 2016-08-16 success 2016-08-17 success 2016-08-18 success 2016-08-19 success 2016-08-22 success 2016-08-23 success 2016-08-24 success 2016-08-25 success 2016-08-26 success 2016-08-29 success 2016-08-30 success 2016-08-31 success 2016-09-01 success 2016-09-02 success 2016-09-05 success 2016-09-06 success 2016-09-07 success 2016-09-08 success 2016-09-09 success 2016-09-12 success 2016-09-13 success 2016-09-14 success 2016-09-19 success 2016-09-20 success 2016-09-21 success 2016-09-22 success 2016-09-23 success 2016-09-26 success 2016-09-27 success 2016-09-28 success 2016-09-29 success 2016-09-30 success 2016-10-10 success 2016-10-11 success 2016-10-12 success 2016-10-13 success 2016-10-14 success 2016-10-17 success 2016-10-18 success 2016-10-19 success 2016-10-20 success 2016-10-21 success 2016-10-24 success 2016-10-25 success 2016-10-26 success 2016-10-27 success 2016-10-28 success 2016-10-31 success 2016-11-01 success 2016-11-02 success 2016-11-03 success 2016-11-04 success 2016-11-07 success 2016-11-08 success 2016-11-09 success 2016-11-10 success 2016-11-11 success 2016-11-14 success 2016-11-15 success 2016-11-16 success 2016-11-17 success 2016-11-18 success 2016-11-21 success 2016-11-22 success 2016-11-23 success 2016-11-24 success 2016-11-25 success 2016-11-28 success 2016-11-29 success 2016-11-30 success 2016-12-01 success 2016-12-02 success 2016-12-05 success 2016-12-06 success 2016-12-07 success 2016-12-08 success 2016-12-09 success 2016-12-12 success 2016-12-13 success 2016-12-14 success 2016-12-15 success 2016-12-16 success 2016-12-19 success 2016-12-20 success 2016-12-21 success 2016-12-22 success 2016-12-23 success 2016-12-26 success 2016-12-27 success 2016-12-28 success 2016-12-29 success 2016-12-30 success 2017-01-03 success 2017-01-04 success 2017-01-05 success 2017-01-06 success 2017-01-09 success 2017-01-10 success 2017-01-11 success 2017-01-12 success 2017-01-13 success 2017-01-16 success 2017-01-17 success 2017-01-18 success 2017-01-19 success 2017-01-20 success 2017-01-23 success 2017-01-24 success 2017-01-25 success 2017-01-26 success 2017-02-03 success 2017-02-06 success 2017-02-07 success 2017-02-08 success 2017-02-09 success 2017-02-10 success 2017-02-13 success 2017-02-14 success 2017-02-15 success 2017-02-16 success 2017-02-17 success 2017-02-20 success 2017-02-21 success 2017-02-22 success 2017-02-23 success 2017-02-24 success 2017-02-27 success 2017-02-28 success 2017-03-01 success 2017-03-02 success 2017-03-03 success 2017-03-06 success 2017-03-07 success 2017-03-08 success 2017-03-09 success 2017-03-10 success 2017-03-13 success 2017-03-14 success 2017-03-15 success 2017-03-16 success 2017-03-17 success 2017-03-20 success 2017-03-21 success 2017-03-22 success 2017-03-23 success 2017-03-24 success 2017-03-27 success 2017-03-28 success 2017-03-29 success 2017-03-30 success 2017-03-31 success 2017-04-05 success 2017-04-06 success 2017-04-07 success 2017-04-10 success 2017-04-11 success 2017-04-12 success 2017-04-13 success 2017-04-14 success 2017-04-17 success 2017-04-18 success 2017-04-19 success 2017-04-20 success 2017-04-21 success 2017-04-24 success 2017-04-25 success 2017-04-26 success 2017-04-27 success 2017-04-28 success 2017-05-02 success 2017-05-03 success 2017-05-04 success 2017-05-05 success 2017-05-08 success 2017-05-09 success 2017-05-10 success 2017-05-11 success 2017-05-12 success 2017-05-15 success 2017-05-16 success 2017-05-17 success 2017-05-18 success 2017-05-19 success 2017-05-22 success 2017-05-23 success 2017-05-24 success 2017-05-25 success 2017-05-26 success 2017-05-31 success 2017-06-01 success 2017-06-02 success 2017-06-05 success 2017-06-06 success 2017-06-07 success 2017-06-08 success 2017-06-09 success 2017-06-12 success 2017-06-13 success 2017-06-14 success 2017-06-15 success 2017-06-16 success 2017-06-19 success 2017-06-20 success 2017-06-21 success 2017-06-22 success 2017-06-23 success 2017-06-26 success 2017-06-27 success 2017-06-28 success 2017-06-29 success 2017-06-30 success 2017-07-03 success 2017-07-04 success 2017-07-05 success 2017-07-06 success 2017-07-07 success 2017-07-10 success 2017-07-11 success 2017-07-12 success 2017-07-13 success 2017-07-14 success 2017-07-17 success 2017-07-18 success 2017-07-19 success 2017-07-20 success 2017-07-21 success 2017-07-24 success 2017-07-25 success 2017-07-26 success 2017-07-27 success 2017-07-28 success 2017-07-31 success 2017-08-01 success 2017-08-02 success 2017-08-03 success 2017-08-04 success 2017-08-07 success 2017-08-08 success 2017-08-09 success 2017-08-10 success 2017-08-11 success 2017-08-14 success 2017-08-15 success 2017-08-16 success 2017-08-17 success 2017-08-18 success 2017-08-21 success 2017-08-22 success 2017-08-23 success 2017-08-24 success 2017-08-25 success 2017-08-28 success 2017-08-29 success 2017-08-30 success 2017-08-31 success 2017-09-01 success 2017-09-04 success 2017-09-05 success 2017-09-06 success 2017-09-07 success 2017-09-08 success 2017-09-11 success 2017-09-12 success 2017-09-13 success 2017-09-14 success 2017-09-15 success 2017-09-18 success 2017-09-19 success 2017-09-20 success 2017-09-21 success 2017-09-22 success 2017-09-25 success 2017-09-26 success 2017-09-27 success 2017-09-28 success 2017-09-29 success 2017-10-09 success 2017-10-10 success 2017-10-11 success 2017-10-12 success 2017-10-13 success 2017-10-16 success 2017-10-17 success 2017-10-18 success 2017-10-19 success 2017-10-20 success 2017-10-23 success 2017-10-24 success 2017-10-25 success 2017-10-26 success 2017-10-27 success 2017-10-30 success 2017-10-31 success 2017-11-01 success 2017-11-02 success 2017-11-03 success 2017-11-06 success 2017-11-07 success 2017-11-08 success 2017-11-09 success 2017-11-10 success 2017-11-13 success 2017-11-14 success 2017-11-15 success 2017-11-16 success 2017-11-17 success 2017-11-20 success 2017-11-21 success 2017-11-22 success 2017-11-23 success 2017-11-24 success 2017-11-27 success 2017-11-28 success 2017-11-29 success 2017-11-30 success 2017-12-01 success 2017-12-04 success 2017-12-05 success 2017-12-06 success 2017-12-07 success 2017-12-08 success 2017-12-11 success 2017-12-12 success 2017-12-13 success 2017-12-14 success 2017-12-15 success 2017-12-18 success 2017-12-19 success 2017-12-20 success 2017-12-21 success 2017-12-22 success 2017-12-25 success 2017-12-26 success 2017-12-27 success 2017-12-28 success 2017-12-29 success 2018-01-02 success 2018-01-03 success 2018-01-04 success 2018-01-05 success 2018-01-08 success 2018-01-09 success 2018-01-10 success 2018-01-11 success 2018-01-12 success 2018-01-15 success 2018-01-16 success 2018-01-17 success 2018-01-18 success 2018-01-19 success 2018-01-22 success 2018-01-23 success 2018-01-24 success 2018-01-25 success 2018-01-26 success 2018-01-29 success 2018-01-30 success 2018-01-31 success 2018-02-01 success 2018-02-02 success 2018-02-05 success 2018-02-06 success 2018-02-07 success 2018-02-08 success 2018-02-09 success 2018-02-12 success 2018-02-13 success 2018-02-14 success 2018-02-22 success 2018-02-23 success 2018-02-26 success 2018-02-27 success 2018-02-28 success 2018-03-01 success 2018-03-02 success 2018-03-05 success 2018-03-06 success 2018-03-07 success 2018-03-08 success 2018-03-09 success 2018-03-12 success 2018-03-13 success 2018-03-14 success 2018-03-15 success 2018-03-16 success 2018-03-19 success 2018-03-20 success 2018-03-21 success 2018-03-22 success 2018-03-23 success 2018-03-26 success 2018-03-27 success 2018-03-28 success 2018-03-29 success 2018-03-30 success 2018-04-02 success 2018-04-03 success 2018-04-04 success 2018-04-09 success 2018-04-10 success 2018-04-11 success 2018-04-12 success 2018-04-13 success 2018-04-16 success 2018-04-17 success 2018-04-18 success 2018-04-19 success 2018-04-20 success 2018-04-23 success 2018-04-24 success 2018-04-25 success 2018-04-26 success 2018-04-27 success 2018-05-02 success 2018-05-03 success 2018-05-04 success 2018-05-07 success 2018-05-08 success 2018-05-09 success 2018-05-10 success 2018-05-11 success 2018-05-14 success 2018-05-15 success 2018-05-16 success 2018-05-17 success 2018-05-18 success 2018-05-21 success 2018-05-22 success 2018-05-23 success 2018-05-24 success 2018-05-25 success 2018-05-28 success 2018-05-29 success 2018-05-30 success 2018-05-31 success 2018-06-01 success 2018-06-04 success 2018-06-05 success 2018-06-06 success 2018-06-07 success 2018-06-08 success 2018-06-11 success 2018-06-12 success 2018-06-13 success 2018-06-14 success 2018-06-15 success 2018-06-19 success 2018-06-20 success 2018-06-21 success 2018-06-22 success 2018-06-25 success 2018-06-26 success 2018-06-27 success 2018-06-28 success 2018-06-29 success 2018-07-02 success 2018-07-03 success 2018-07-04 success 2018-07-05 success 2018-07-06 success 2018-07-09 success 2018-07-10 success 2018-07-11 success 2018-07-12 success 2018-07-13 success 2018-07-16 success 2018-07-17 success 2018-07-18 success 2018-07-19 success 2018-07-20 success 2018-07-23 success 2018-07-24 success 2018-07-25 success 2018-07-26 success 2018-07-27 success 2018-07-30 success 2018-07-31 success 2018-08-01 success 2018-08-02 success 2018-08-03 success 2018-08-06 success 2018-08-07 success 2018-08-08 success 2018-08-09 success 2018-08-10 success 2018-08-13 success 2018-08-14 success 2018-08-15 success 2018-08-16 success 2018-08-17 success 2018-08-20 success 2018-08-21 success 2018-08-22 success 2018-08-23 success 2018-08-24 success 2018-08-27 success 2018-08-28 success 2018-08-29 success 2018-08-30 success 2018-08-31 success 2018-09-03 success 2018-09-04 success 2018-09-05 success 2018-09-06 success 2018-09-07 success 2018-09-10 success 2018-09-11 success 2018-09-12 success 2018-09-13 success 2018-09-14 success 2018-09-17 success 2018-09-18 success 2018-09-19 success 2018-09-20 success 2018-09-21 success 2018-09-25 success 2018-09-26 success 2018-09-27 success 2018-09-28 success 2018-10-08 success 2018-10-09 success 2018-10-10 success 2018-10-11 success 2018-10-12 success 2018-10-15 success 2018-10-16 success 2018-10-17 success 2018-10-18 success 2018-10-19 success 2018-10-22 success 2018-10-23 success 2018-10-24 success 2018-10-25 success 2018-10-26 success 2018-10-29 success 2018-10-30 success 2018-10-31 success 2018-11-01 success 2018-11-02 success 2018-11-05 success 2018-11-06 success 2018-11-07 success 2018-11-08 success 2018-11-09 success 2018-11-12 success 2018-11-13 success 2018-11-14 success 2018-11-15 success 2018-11-16 success 2018-11-19 success 2018-11-20 success 2018-11-21 success 2018-11-22 success 2018-11-23 success 2018-11-26 success 2018-11-27 success 2018-11-28 success 2018-11-29 success 2018-11-30 success 2018-12-03 success 2018-12-04 success 2018-12-05 success 2018-12-06 success 2018-12-07 success 2018-12-10 success 2018-12-11 success 2018-12-12 success 2018-12-13 success 2018-12-14 success 2018-12-17 success 2018-12-18 success 2018-12-19 success 2018-12-20 success 2018-12-21 success 2018-12-24 success 2018-12-25 success 2018-12-26 success 2018-12-27 success 2018-12-28 success 2019-01-02 success 2019-01-03 success 2019-01-04 success 2019-01-07 success 2019-01-08 success 2019-01-09 success 2019-01-10 success 2019-01-11 success 2019-01-14 success 2019-01-15 success 2019-01-16 success 2019-01-17 success 2019-01-18 success 2019-01-21 success 2019-01-22 success 2019-01-23 success 2019-01-24 success 2019-01-25 success 2019-01-28 success 2019-01-29 success 2019-01-30 success 2019-01-31 success 2019-02-01 success 2019-02-11 success 2019-02-12 success 2019-02-13 success 2019-02-14 success 2019-02-15 success 2019-02-18 success 2019-02-19 success 2019-02-20 success 2019-02-21 success 2019-02-22 success 2019-02-25 success 2019-02-26 success 2019-02-27 success 2019-02-28 success 2019-03-01 success 2019-03-04 success 2019-03-05 success 2019-03-06 success 2019-03-07 success 2019-03-08 success 2019-03-11 success 2019-03-12 success 2019-03-13 success 2019-03-14 success 2019-03-15 success 2019-03-18 success 2019-03-19 success 2019-03-20 success 2019-03-21 success 2019-03-22 success 2019-03-25 success 2019-03-26 success 2019-03-27 success 2019-03-28 success 2019-03-29 success 2019-04-01 success 2019-04-02 success 2019-04-03 success 2019-04-04 success 2019-04-08 success 2019-04-09 success 2019-04-10 success 2019-04-11 success 2019-04-12 success 2019-04-15 success 2019-04-16 success 2019-04-17 success 2019-04-18 success 2019-04-19 success 2019-04-22 success 2019-04-23 success 2019-04-24 success 2019-04-25 success 2019-04-26 success 2019-04-29 success 2019-04-30 success 2019-05-06 success 2019-05-07 success 2019-05-08 success 2019-05-09 success 2019-05-10 success 2019-05-13 success 2019-05-14 success 2019-05-15 success 2019-05-16 success 2019-05-17 success 2019-05-20 success 2019-05-21 success 2019-05-22 success 2019-05-23 success 2019-05-24 success 2019-05-27 success 2019-05-28 success 2019-05-29 success 2019-05-30 success 2019-05-31 success 2019-06-03 success 2019-06-04 success 2019-06-05 success 2019-06-06 success 2019-06-10 success 2019-06-11 success 2019-06-12 success 2019-06-13 success 2019-06-14 success 2019-06-17 success 2019-06-18 success 2019-06-19 success 2019-06-20 success 2019-06-21 success 2019-06-24 success 2019-06-25 success 2019-06-26 success 2019-06-27 success 2019-06-28 success 2019-07-01 success 2019-07-02 success 2019-07-03 success 2019-07-04 success 2019-07-05 success 2019-07-08 success 2019-07-09 success 2019-07-10 success 2019-07-11 success 2019-07-12 success 2019-07-15 success 2019-07-16 success 2019-07-17 success 2019-07-18 success 2019-07-19 success 2019-07-22 success 2019-07-23 success 2019-07-24 success 2019-07-25 success 2019-07-26 success 2019-07-29 success 2019-07-30 success 2019-07-31 success 2019-08-01 success 2019-08-02 success 2019-08-05 success 2019-08-06 success 2019-08-07 success 2019-08-08 success 2019-08-09 success 2019-08-12 success 2019-08-13 success 2019-08-14 success 2019-08-15 success 2019-08-16 success 2019-08-19 success 2019-08-20 success 2019-08-21 success 2019-08-22 success 2019-08-23 success 2019-08-26 success 2019-08-27 success 2019-08-28 success 2019-08-29 success 2019-08-30 以储存数据：zbzs_datas.pkl
# 读取数据pkl_file = open('hs300_datas.pkl', 'rb')hs300_datas = pickle.load(pkl_file)pkl_file = open('zz500_datas.pkl', 'rb')zz500_datas = pickle.load(pkl_file)pkl_file = open('cyb_datas.pkl', 'rb')cyb_datas = pickle.load(pkl_file)pkl_file = open('sz50_datas.pkl', 'rb')sz50_datas = pickle.load(pkl_file)pkl_file = open('szzs_datas.pkl', 'rb')szzs_datas = pickle.load(pkl_file)pkl_file = open('zbzs_datas.pkl', 'rb')zbzs_datas = pickle.load(pkl_file)# 根据日收益率计算信号def daily_factor(df_dic,index_id):# 计算信号hs300_close=df_dic['close']start=min(hs300_close.index)end=max(hs300_close.index)index_close=get_price(index_id,start_date=start,end_date=end,fields='close')hs300_ret=hs300_close.pct_change()index_ret=index_close.pct_change()index_mean_ret=index_ret.rolling(21).mean()index_mean_ret=index_mean_ret[21:]hs300_ret=hs300_ret[1:]index_ret=index_ret[1:]excess_ret=abs(hs300_ret-np.broadcast_to(index_ret,hs300_ret.shape))csad=excess_ret[~hs300_close.isna()].mean(axis=1)temp=[]for i in hs300_ret.index[21:]:x=hs300_ret.loc[i].dropna().valuesX=np.column_stack((x, x**2))y=np.broadcast_to(csad.loc[i],(len(x),))beta=np.linalg.lstsq(X,y)[0][1] # r_2系数temp.append([i,beta])df=pd.DataFrame(temp,columns=['date','beta']).set_index('date')df=df.join(index_close)df['index_mean_ret']=df.close.pct_change().rolling(21).mean()df['r_up_singal']=(df.beta<0)&(df.index_mean_ret>0)df=df[21:]# 计算scorescore_temp=[]for idx,row in df.iterrows():score_temp.append(cal_score(row['beta'],row['index_mean_ret']))df['r_factor']=score_tempreturn df根据日收益率计算信号并在每个指数上回测观察其表现，可以看到日收益的表现不及月度收益率（月度收益率见：羊群效应策略表现附表)
回测方法：当信号出现时持有，信号消失后平仓
信号：$CSAD<0\ and\ 指数平均收益率>0$
由下表可以看到除在上证50上稳定性较差外，在其他指标上的收益都较为稳定均大于等于1
# 上涨=>查看按照研报信号各宽基指数风险指标datas = [hs300_datas, zz500_datas, cyb_datas, sz50_datas,szzs_datas,zbzs_datas]index_name = ['000300.XSHG', '000905.XSHG', '399006.XSHE', '000016.XSHG','000001.XSHG','399101.XSHE']index_data_dic=dict(zip(index_name,datas))# 储存回测风险指标report_df = pd.DataFrame()# 储存净值用于作图net_dic={}for index_data, n in zip(datas, index_name):factor_df = daily_factor(index_data, n)# 滞后一期的收益ret_arr = factor_df['close'].pct_change().shift(-1)ret_arr = np.nan_to_num(ret_arr.values)position = factor_df.r_up_singal.values * np.ones(len(factor_df))ret = position * ret_arr # 收益率cum_ret = (1 + ret).cumprod() # 净值# 最大回撤max_nv = np.maximum.accumulate(cum_ret)mdd = -np.min(cum_ret / max_nv - 1)# 胜率winning_ratio = np.sum(np.where(ret > 0, 1, 0)) / np.sum(position)# 计算年化收益率annual_ret = cum_ret[-1]**(240 / (len(ret) - 5)) - 1# 计算累计收益率cum_ret_rate = cum_ret[-1] - 1# 夏普sharpe = ret.mean() / ret.std() * np.sqrt(240)df = pd.DataFrame({'年化收益率': '{:.2%}'.format(annual_ret),'累计收益率': '{:.2%}'.format(cum_ret_rate),'胜率': '{:.2%}'.format(winning_ratio),'开仓天数': np.sum(position),'最大回撤': '{:.2%}'.format(mdd),'夏普': '{:.2}'.format(sharpe)},index=[n])report_df = report_df.append(df)# 作图基准使用hs300if n=='000300.XSHG':net_dic['benchmark']=factor_df['close']/factor_df['close'][0]# 储存净值曲线net_dic[n]=cum_ret# 查看风险指标print('各指数风险指标:')display(HTML(report_df.to_html()))nev_df=pd.DataFrame(net_dic)# 画图plt.figure()fig = plt.figure(figsize=(20, 10))ax1 = fig.add_subplot(1, 1, 1)for i in nev_df.columns:ax1.plot(nev_df[i], label=i)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('时间')plt.ylabel('净值')plt.title('各组净值曲线')plt.show()各指数风险指标:
| 年化收益率 | 累计收益率 | 胜率 | 开仓天数 | 最大回撤 | 夏普 | |
|---|---|---|---|---|---|---|
| 000300.XSHG | 8.54% | 58.92% | 60.00% | 200.0 | 10.25% | 1.0 |
| 000905.XSHG | 14.91% | 119.44% | 64.22% | 204.0 | 11.34% | 1.6 |
| 399006.XSHE | 13.08% | 100.37% | 56.00% | 200.0 | 24.34% | 1.1 |
| 000016.XSHG | 7.81% | 53.00% | 53.41% | 249.0 | 20.16% | 0.85 |
| 000001.XSHG | 12.59% | 95.56% | 65.15% | 198.0 | 7.70% | 1.5 |
| 399101.XSHE | 16.80% | 140.63% | 66.85% | 181.0 | 9.11% | 1.9 |
<Figure size 432x288 with 0 Axes>
在沪深300上测试日度收益和月度收益计算信号指标的风险指标，比较分析风险指标情况：
可以看日度收益率计算的信号出现后1至3日夏普及收益率较好，其中持有3日年化收益率为17.41%，夏普为1.3
# index_ret为22日收益率平均值度，csad为日收益率计算df=daily_factor(hs300_datas, '000300.XSHG')temp = []params = 'ret'holding = [1, 2, 3, 5, 10, 15, 20,22, 25, 30]for j in holding:bt = back_test(df, 'r_up_singal', j)cum_ret = bt['cum_ret']ret = bt['ret']temp.append(['{:.2%}'.format(cum_ret[-1]**(240 / (len(cum_ret) - 5)) - 1),'{:.2}'.format(ret.mean() / ret.std() * np.sqrt(240))])temp = np.array(temp)columns_name = list(map(lambda x: '持有{}天'.format(x), holding))temp_df = pd.DataFrame(temp,columns=['ret','sharpe'],index=columns_name).Tprint('日度收益率信号指标')temp_df日度收益率信号指标
.dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; }
| 持有1天 | 持有2天 | 持有3天 | 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有22天 | 持有25天 | 持有30天 | |
|---|---|---|---|---|---|---|---|---|---|---|
| ret | 8.54% | 13.38% | 17.41% | 14.48% | 10.62% | 10.48% | 6.94% | 4.04% | 3.72% | 6.10% |
| sharpe | 1.0 | 1.2 | 1.3 | 0.96 | 0.64 | 0.6 | 0.42 | 0.29 | 0.28 | 0.37 |
月度收益计算的信号再有持有期时均**不及**日度收益率计算的信号
# index_ret为月收益率信号，csad为月收益率计算factor_hs300=get_factor(hs300_datas, '000300.XSHG',22)temp = []params = 'ret'holding = [1, 2, 3, 5, 10, 15, 20,22, 25, 30]for j in holding:bt = back_test(factor_hs300, 'r_up_singal', j)cum_ret = bt['cum_ret']ret = bt['ret']temp.append(['{:.2%}'.format(cum_ret[-1]**(240 / (len(cum_ret) - 5)) - 1),'{:.2}'.format(ret.mean() / ret.std() * np.sqrt(240))])temp = np.array(temp)columns_name = list(map(lambda x: '持有{}天'.format(x), holding))temp_df = pd.DataFrame(temp,columns=['ret','sharpe'],index=columns_name).Tprint('月度收益率信号指标')temp_df月度收益率信号指标
.dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; }
| 持有1天 | 持有2天 | 持有3天 | 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有22天 | 持有25天 | 持有30天 | |
|---|---|---|---|---|---|---|---|---|---|---|
| ret | 12.82% | 13.23% | 14.84% | 9.66% | 5.11% | 7.05% | 6.12% | 7.38% | 8.59% | 7.95% |
| sharpe | 0.83 | 0.83 | 0.89 | 0.59 | 0.34 | 0.42 | 0.38 | 0.43 | 0.47 | 0.44 |
结论：通过在沪深300上测试发现，在有持有期的情况下日度收益率计算的信号时优于月度收益计算的信号指标；在无持有期的情况下，月度指标要由于日度指标。
日度收益率计算指标在持有期为3日时的净值情况及风险指标
# up_singal效果比mean_singal好，holding=3时up_singal比较强back_test(df,'r_up_singal',3)summary(df)
年化收益率: 17.41% 累计收益率: 147.83% 最大回撤: 10.96% 夏普比率：1.3
按日度益率计算CSAD,在根据22日的指数收益率平均值确认时上涨/下跌
我们根据**CSAD**及**指数22日平均收益率收益率**数值计算score，并将score从大到小分为五组进行回测；
score的构建：
# 5-1 将CSAD和指数收益*打分def cal_score(csad, ret):if csad < 0 and ret > 0:score = csad + retelse:score=abs(csad)+abs(ret)return score
可以看到日度收益率在短期(3日)持有使由较好的收益，及较为稳定的夏普比率
# 日度收益率计算信号指标风险指标情况df=daily_factor(index_data_dic['000300.XSHG'],'000300.XSHG')# 夏普敏感性分析sharpe_df = threshold_analysis(df, threshold_col='r_factor', holding=[3,5,10,15,20], params='sharpe')# 收益率敏感性分析ret_df = threshold_analysis(df,threshold_col='r_factor',holding=[3,5,10,15,20],params='ret')print('夏普敏感性分析:')display(HTML(sharpe_df.to_html()))print('收率敏感性分析:')display(HTML(ret_df.to_html()))夏普敏感性分析:
| 持有3天 | 持有5天 | 持有10天 | 持有15天 | 持有20天 | |
|---|---|---|---|---|---|
| 273.2 | 0.557225 | 0.520973 | 0.520973 | 0.520973 | 0.520973 |
| 545.4 | 0.637869 | 0.505993 | 0.520973 | 0.520973 | 0.520973 |
| 817.6 | 0.594503 | 0.517535 | 0.486738 | 0.520577 | 0.520973 |
| 1089.8 | 0.965465 | 0.348784 | 0.323627 | 0.411863 | 0.480606 |
收率敏感性分析:
| 持有3天 | 持有5天 | 持有10天 | 持有15天 | 持有20天 | |
|---|---|---|---|---|---|
| 273.2 | 10.96% | 10.04% | 10.04% | 10.04% | 10.04% |
| 545.4 | 12.77% | 9.64% | 10.04% | 10.04% | 10.04% |
| 817.6 | 10.81% | 9.67% | 9.11% | 10.02% | 10.04% |
| 1089.8 | 15.48% | 4.95% | 4.84% | 7.09% | 8.87% |
按月度收益率计算CSAD,根据月度收益率确认上涨/下跌
由下表可以看出在持有期为15日时月度收益率信号较好
# 月度收益率计算信号指标风险指标情况# 夏普敏感性分析factor_hs300 = get_factor(hs300_datas, '000300.XSHG', 22)sharpe_df = threshold_analysis(factor_hs300, threshold_col='r_factor', holding=[3,5,10,15,20], params='sharpe')# 收益率敏感性分析ret_df = threshold_analysis(factor_hs300,threshold_col='r_factor',holding=[3,5,10,15,20],params='ret')print('夏普敏感性分析:')display(HTML(sharpe_df.to_html()))print('收率敏感性分析:')display(HTML(ret_df.to_html()))夏普敏感性分析:
| 持有3天 | 持有5天 | 持有10天 | 持有15天 | 持有20天 | |
|---|---|---|---|---|---|
| 277.6 | 0.518417 | 0.383144 | 0.472864 | 0.496967 | 0.489153 |
| 554.2 | 0.520350 | 0.255122 | 0.326271 | 0.411510 | 0.404458 |
| 830.8 | 0.873920 | 0.716314 | 0.911090 | 1.024472 | 0.918649 |
| 1107.4 | 0.012563 | -0.022069 | 0.323722 | 0.418821 | 0.464447 |
收率敏感性分析:
| 持有3天 | 持有5天 | 持有10天 | 持有15天 | 持有20天 | |
|---|---|---|---|---|---|
| 277.6 | 9.46% | 6.35% | 8.67% | 9.33% | 9.14% |
| 554.2 | 8.27% | 3.15% | 4.88% | 6.99% | 6.93% |
| 830.8 | 10.49% | 9.61% | 14.28% | 17.58% | 16.22% |
| 1107.4 | -0.32% | -0.82% | 3.31% | 4.88% | 5.86% |
结论：通过在沪深300上的测试，可以发现在构建SCORE后月度收益率后续的可提升空间比日度收益率构建的SCORE提升空间更大,***后续我们通过月度收益计算信号指标进行分析***。
我们在六种宽基指数：上证综指、上证 50、沪深 300、中证 500、中小板综、创业板指，通过 CCK 模型判断指数成分股 间是否存在羊群效应，在羊群效应发生后买入/卖出。
策略步骤：计算向前 22 日（包括当日）每天的成分股组合截面绝对离 散度 CSAD，OLS 估计 CCK 模型中$R_{m,t}^2$的系数$R^2$，若$R^2$显著为负则认为当日该组合存在羊群效应，根据 22 日内指数平均收益率的正负区分羊群效应发生时的市场趋势为上涨/下跌，买入/卖出标的指数并持仓22 交易日，持有期不重复开仓。
ps：过滤指数中上市不足三月\st\当日停牌股票
回测实际范围：2014.1.1至2019.8.1
回测方法：当$R^2$的显著为负则认为当日该组合存在羊群效应，$R_m$为根据 22 日指数收益率的正负区分羊群效应发生时的市场趋势为上涨/下跌，既$R^2$为负，$R_m$大于零时，开仓；不满足此条件时平仓或不开仓。
市场趋势为上涨时策略在宽基指数上的平均表现如下所示：
可以看到羊群效应在创业版及中小企业板块中效果较好。
各指指数大体情况如下：
**创业板指数**年化收益为18%，夏普为0.9，胜率51.77%;
**中小企业综指**年化收益为17.58%，夏普为1，胜率55.98%;
**中证500**年化收益为15.49%，夏普为0.9，胜率56.08%;
**上证50**回测效果最差，年化收益为11.5%，夏普为0.74，胜率51.29%;
# 上涨=>查看按照研报信号各宽基指数风险指标datas = [hs300_datas, zz500_datas, cyb_datas, sz50_datas,szzs_datas,zbzs_datas]index_name = ['000300.XSHG', '000905.XSHG', '399006.XSHE', '000016.XSHG','000001.XSHG','399101.XSHE']# 储存回测风险指标report_df = pd.DataFrame()# 储存净值用于作图net_dic={}for index_data, n in zip(datas, index_name):factor_df = get_factor(index_data, n, 22)# 滞后一期的收益ret_arr = factor_df['close'].pct_change().shift(-1)ret_arr = np.nan_to_num(ret_arr.values)position = factor_df.r_up_singal.values * np.ones(len(factor_df))ret = position * ret_arr # 收益率cum_ret = (1 + ret).cumprod() # 净值# 最大回撤max_nv = np.maximum.accumulate(cum_ret)mdd = -np.min(cum_ret / max_nv - 1)# 胜率winning_ratio = np.sum(np.where(ret > 0, 1, 0)) / np.sum(position)# 计算年化收益率annual_ret = cum_ret[-1]**(240 / (len(ret) - 5)) - 1# 计算累计收益率cum_ret_rate = cum_ret[-1] - 1# 夏普sharpe = ret.mean() / ret.std() * np.sqrt(240)df = pd.DataFrame({'年化收益率': '{:.2%}'.format(annual_ret),'累计收益率': '{:.2%}'.format(cum_ret_rate),'胜率': '{:.2%}'.format(winning_ratio),'开仓天数': np.sum(position),'最大回撤': '{:.2%}'.format(mdd),'夏普': '{:.2}'.format(sharpe)},index=[n])report_df = report_df.append(df)# 作图基准使用hs300if n=='000300.XSHG':net_dic['benchmark']=factor_df['close']/factor_df['close'][0]# 储存净值曲线net_dic[n]=cum_ret# 查看风险指标print('各指数风险指标:')display(HTML(report_df.to_html()))nev_df=pd.DataFrame(net_dic)# 画图plt.figure()fig = plt.figure(figsize=(18, 8))ax1 = fig.add_subplot(1, 1, 1)for i in nev_df.columns:ax1.plot(nev_df[i], label=i)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('时间')plt.ylabel('净值')plt.title('各组净值曲线')plt.show()各指数风险指标:
| 年化收益率 | 累计收益率 | 胜率 | 开仓天数 | 最大回撤 | 夏普 | |
|---|---|---|---|---|---|---|
| 000300.XSHG | 12.82% | 99.94% | 54.21% | 795.0 | 28.07% | 0.83 |
| 000905.XSHG | 15.49% | 128.72% | 56.08% | 765.0 | 35.59% | 0.9 |
| 399006.XSHE | 18.00% | 158.85% | 51.77% | 678.0 | 40.42% | 0.9 |
| 000016.XSHG | 11.50% | 86.92% | 51.29% | 774.0 | 24.33% | 0.74 |
| 000001.XSHG | 12.41% | 95.84% | 56.41% | 780.0 | 27.01% | 0.83 |
| 399101.XSHE | 17.58% | 153.53% | 55.98% | 761.0 | 33.07% | 1.0 |
<Figure size 432x288 with 0 Axes>
市场下跌阶段，创业板的效果依旧明显，具体情况如下表,
# 下跌=>查看按照研报信号各宽基指数风险指标datas = [hs300_datas, zz500_datas, cyb_datas, sz50_datas,szzs_datas,zbzs_datas]index_name = ['000300.XSHG', '000905.XSHG', '399006.XSHE', '000016.XSHG','000001.XSHG','399101.XSHE']# 储存回测风险指标report_df = pd.DataFrame()# 储存净值用于作图net_dic={}for index_data, n in zip(datas, index_name):factor_df = get_factor(index_data, n, 22)# 滞后一期的收益ret_arr = factor_df['close'].pct_change().shift(-1)ret_arr = np.nan_to_num(ret_arr.values)position = factor_df.r_down_singal.values * np.ones(len(factor_df))ret = position * ret_arr # 收益率cum_ret = (1 + ret).cumprod() # 净值# 最大回撤max_nv = np.maximum.accumulate(cum_ret)mdd = -np.min(cum_ret / max_nv - 1)# 胜率winning_ratio = np.sum(np.where(ret > 0, 1, 0)) / np.sum(position)# 计算年化收益率annual_ret = cum_ret[-1]**(240 / (len(ret) - 5)) - 1# 计算累计收益率cum_ret_rate = cum_ret[-1] - 1# 夏普sharpe = ret.mean() / ret.std() * np.sqrt(240)df = pd.DataFrame({'年化收益率': '{:.2%}'.format(annual_ret),'累计收益率': '{:.2%}'.format(cum_ret_rate),'胜率': '{:.2%}'.format(winning_ratio),'开仓天数': np.sum(position),'最大回撤': '{:.2%}'.format(mdd),'夏普': '{:.2}'.format(sharpe)},index=[n])report_df = report_df.append(df)# 作图基准使用hs300if n=='000300.XSHG':net_dic['benchmark']=factor_df['close']/factor_df['close'][0]# 储存净值曲线net_dic[n]=cum_ret# 查看风险指标print('各指数风险指标:')from IPython.core.display import HTMLdisplay(HTML(report_df.to_html()))nev_df=pd.DataFrame(net_dic)# 画图plt.figure()fig = plt.figure(figsize=(18, 8))ax1 = fig.add_subplot(1, 1, 1)for i in nev_df.columns:ax1.plot(nev_df[i], label=i)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('时间')plt.ylabel('净值')plt.title('各组净值曲线')plt.show()各指数风险指标:
| 年化收益率 | 累计收益率 | 胜率 | 开仓天数 | 最大回撤 | 夏普 | |
|---|---|---|---|---|---|---|
| 000300.XSHG | -3.49% | -18.44% | 50.42% | 589.0 | 41.87% | -0.11 |
| 000905.XSHG | -9.66% | -44.20% | 52.02% | 619.0 | 56.68% | -0.37 |
| 399006.XSHE | -12.08% | -52.29% | 49.29% | 706.0 | 62.81% | -0.42 |
| 000016.XSHG | -0.24% | -1.38% | 50.90% | 609.0 | 38.00% | 0.075 |
| 000001.XSHG | -6.10% | -30.35% | 51.32% | 604.0 | 44.84% | -0.29 |
| 399101.XSHE | -9.36% | -43.15% | 50.72% | 623.0 | 53.43% | -0.38 |
<Figure size 432x288 with 0 Axes>
我们根据CSAD及指数收益率数值计算score，并将score从大到小分为五组进行回测；
score的构建：
# 5-1 将CSAD和指数收益*打分def cal_score(csad, ret):if csad < 0 and ret > 0:score = csad + retelse:score=abs(csad)+abs(ret)return score
# 获取因子factor_hs300 = get_factor(hs300_datas, '000300.XSHG', 22)# 分组group_hs = get_group(factor_hs300, 5, 'r_factor')group_hs['未来5日涨幅'] = group_hs.close.pct_change(4).shift(-4)group_hs['未来15日涨幅'] = group_hs.close.pct_change(14).shift(-14)group_hs['未来20日涨幅'] = group_hs.close.pct_change(19).shift(-19)
由下图可以看到第4组在当日、未来5日、未来15日、未来20日的收益都较为稳定
# 查看因子与未来收益率的关系def group_plot(df, y):a = df[df['GROUP'] != 0]b = a.groupby('GROUP').mean()plt.xlabel('CSAD')plt.ylabel(y)plt.bar(np.array(b.index), b[y], width=0.5)plt.figure(1, figsize=(12, 6))plt.title('当日涨幅')group_plot(group_hs, '当日涨幅')plt.figure(2, figsize=(12, 6))plt.title('未来5日涨幅')group_plot(group_hs, '未来5日涨幅')plt.figure(3, figsize=(12, 6))plt.title('未来15日涨幅')group_plot(group_hs, '未来15日涨幅')plt.figure(4, figsize=(12, 6))plt.title('未来20日涨幅')group_plot(group_hs, '未来20日涨幅')根据沪深300的score进行分组回测，持有天数为22日
# 获取分组回测数据cum_ret_dic, report = group_back_test(group_hs, 'GROUP', 22)# 构建dfcum_df = pd.DataFrame(cum_ret_dic, index=factor_hs300.index[:-1])# 获取分组名称group_list = cum_df.columns# 画图plt.figure()fig = plt.figure(figsize=(18, 8))ax1 = fig.add_subplot(1, 1, 1)for i in group_list:ax1.plot(cum_df[i], label=i)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('时间')plt.ylabel('净值')plt.title('各组净值曲线')plt.show()<Figure size 432x288 with 0 Axes>
# 多空收益d = cum_df['G5'] - cum_df['G1']plt.figure()fig = plt.figure(figsize=(13, 5))ax1 = fig.add_subplot(1, 1, 1)ax1.plot(d, label='多组合')ax1.axhline(1.0, linestyle='-', color='black', lw=1)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('time')plt.ylabel('net value ratio')plt.title('多空净值组合')plt.show()<Figure size 432x288 with 0 Axes>
第四组年化收益率为15.13%，夏普为0.91较为稳定
# 回测报告pd.DataFrame(report).T
.dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; }
| 夏普比率 | 年化收益率 | 日胜率 | 最大回撤 | 满仓天数 | 空仓天数 | 累计收益率 | |
|---|---|---|---|---|---|---|---|
| G1 | 0.32 | 3.69% | 50.77% | 29.03% | 776 | 608 | 23.15% |
| G2 | 0.48 | 7.52% | 52.75% | 33.14% | 910 | 474 | 51.66% |
| G3 | 0.32 | 4.49% | 54.22% | 51.56% | 699 | 685 | 28.72% |
| G4 | 0.91 | 15.13% | 53.45% | 31.86% | 999 | 385 | 124.54% |
| G5 | 0.49 | 6.33% | 51.24% | 37.43% | 927 | 457 | 42.23% |
通过敏感性分析我们可以看到在持有天数为15天时,夏普为**1.02**年化收益率为**17.58%**达到最优
# 夏普敏感性分析sharpe_df = threshold_analysis(factor_hs300, threshold_col='r_factor', params='sharpe')# 收益率敏感性分析ret_df = threshold_analysis(factor_hs300,threshold_col='r_factor',params='ret')print('夏普敏感性分析:')display(HTML(sharpe_df.to_html()))print('收率敏感性分析:')display(HTML(ret_df.to_html()))夏普敏感性分析:
| 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有25天 | |
|---|---|---|---|---|---|
| 277.6 | 0.383144 | 0.472864 | 0.496967 | 0.489153 | 0.484300 |
| 554.2 | 0.255122 | 0.326271 | 0.411510 | 0.404458 | 0.424605 |
| 830.8 | 0.716314 | 0.911090 | 1.024472 | 0.918649 | 0.761617 |
| 1107.4 | -0.022069 | 0.323722 | 0.418821 | 0.464447 | 0.589634 |
收率敏感性分析:
| 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有25天 | |
|---|---|---|---|---|---|
| 277.6 | 6.35% | 8.67% | 9.33% | 9.14% | 9.03% |
| 554.2 | 3.15% | 4.88% | 6.99% | 6.93% | 7.48% |
| 830.8 | 9.61% | 14.28% | 17.58% | 16.22% | 14.00% |
| 1107.4 | -0.82% | 3.31% | 4.88% | 5.86% | 8.05% |
# 获取因子factor_zz500 = get_factor(zz500_datas, '000905.XSHG', 22)# 分组group_zz = get_group(factor_zz500, 5, 'r_factor')group_zz['未来5日涨幅'] = group_zz.close.pct_change(4).shift(-4)group_zz['未来15日涨幅'] = group_zz.close.pct_change(14).shift(-14)group_zz['未来20日涨幅'] = group_zz.close.pct_change(19).shift(-19)
与沪深300的情况不同在持有当日与未来5日的各组收益不稳定；G5在未来15、20日收益较高
# 绘图plt.figure(1, figsize=(12, 6))plt.title('当日涨幅')group_plot(group_zz, '当日涨幅')plt.figure(2, figsize=(12, 6))plt.title('未来5日涨幅')group_plot(group_zz, '未来5日涨幅')plt.figure(3, figsize=(12, 6))plt.title('未来15日涨幅')group_plot(group_zz, '未来15日涨幅')plt.figure(4, figsize=(12, 6))plt.title('未来20日涨幅')group_plot(group_zz, '未来20日涨幅')根据中证500的score进行分组回测，持有天数为20日
# 获取分组回测数据cum_ret_dic, report = group_back_test(group_zz, 'GROUP', 20)# 构建dfcum_df = pd.DataFrame(cum_ret_dic, index=factor_zz500.index[:-1])# 获取分组名称group_list = cum_df.columns# 画图plt.figure()fig = plt.figure(figsize=(18, 8))ax1 = fig.add_subplot(1, 1, 1)for i in group_list:ax1.plot(cum_df[i], label=i)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('时间')plt.ylabel('净值')plt.title('各组净值曲线')plt.show()<Figure size 432x288 with 0 Axes>
# 多空收益净值d = cum_df['G5'] - cum_df['G1']plt.figure()fig = plt.figure(figsize=(13, 5))ax1 = fig.add_subplot(1, 1, 1)ax1.plot(d, label='多组合')ax1.axhline(1.0, linestyle='-', color='black', lw=1)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('time')plt.ylabel('net value ratio')plt.title('多空净值组合')plt.show()<Figure size 432x288 with 0 Axes>
从上面的净值曲线及下表的风险指标中可以看到羊群效应在中证500上的择时效果并不理想。各组单调性及稳定性也不理想。
# 回测报告pd.DataFrame(report).T
.dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; }
| 夏普比率 | 年化收益率 | 日胜率 | 最大回撤 | 满仓天数 | 空仓天数 | 累计收益率 | |
|---|---|---|---|---|---|---|---|
| G1 | 0.014 | -2.08% | 52.85% | 44.87% | 772 | 612 | -11.35% |
| G2 | -0.14 | -6.03% | 52.05% | 48.09% | 953 | 431 | -30.03% |
| G3 | 0.011 | -2.22% | 55.87% | 61.56% | 682 | 702 | -12.08% |
| G4 | 0.21 | 2.19% | 55.56% | 54.17% | 837 | 547 | 13.24% |
| G5 | 0.3 | 3.97% | 53.07% | 40.60% | 863 | 521 | 25.04% |
通过敏感性分析我们可以看到在持有天数为10天时,夏普为**0.83**年化收益率为**11.93%**达到最优,但总体效果并不好
# 夏普敏感性分析sharpe_df = threshold_analysis(factor_zz500, threshold_col='r_factor', params='sharpe')# 收益率敏感性分析ret_df = threshold_analysis(factor_zz500,threshold_col='r_factor',params='ret')print('夏普敏感性分析:')display(HTML(sharpe_df.to_html()))print('收率敏感性分析:')display(HTML(ret_df.to_html()))夏普敏感性分析:
| 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有25天 | |
|---|---|---|---|---|---|
| 277.6 | 0.201438 | 0.291525 | 0.235250 | 0.278231 | 0.292724 |
| 554.2 | 0.294454 | 0.134756 | 0.279986 | 0.310918 | 0.248219 |
| 830.8 | -0.012515 | -0.011819 | 0.204884 | 0.305247 | 0.172280 |
| 1107.4 | 0.459823 | 0.831499 | 0.398340 | 0.301817 | 0.237495 |
收率敏感性分析:
| 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有25天 | |
|---|---|---|---|---|---|
| 277.6 | 1.81% | 4.29% | 2.69% | 3.92% | 4.34% |
| 554.2 | 4.15% | 0.28% | 3.93% | 4.81% | 3.06% |
| 830.8 | -2.05% | -2.55% | 2.08% | 4.55% | 1.07% |
| 1107.4 | 5.34% | 11.93% | 5.50% | 3.97% | 2.79% |
# 获取因子factor_sz50 = get_factor(zz500_datas, '000016.XSHG', 22)# 分组group_sz = get_group(factor_sz50, 5, 'r_factor')group_sz['未来5日涨幅'] = group_sz.close.pct_change(4).shift(-4)group_sz['未来15日涨幅'] = group_sz.close.pct_change(14).shift(-14)group_sz['未来20日涨幅'] = group_sz.close.pct_change(19).shift(-19)
# 绘图plt.figure(1, figsize=(12, 6))plt.title('当日涨幅')group_plot(group_sz, '当日涨幅')plt.figure(2, figsize=(12, 6))plt.title('未来5日涨幅')group_plot(group_sz, '未来5日涨幅')plt.figure(3, figsize=(12, 6))plt.title('未来15日涨幅')group_plot(group_sz, '未来15日涨幅')plt.figure(4, figsize=(12, 6))plt.title('未来20日涨幅')group_plot(group_sz, '未来20日涨幅')# 获取分组回测数据cum_ret_dic, report = group_back_test(group_sz, 'GROUP', 20)# 构建dfcum_df = pd.DataFrame(cum_ret_dic, index=factor_zz500.index[:-1])# 获取分组名称group_list = cum_df.columns# 画图plt.figure()fig = plt.figure(figsize=(18, 8))ax1 = fig.add_subplot(1, 1, 1)for i in group_list:ax1.plot(cum_df[i], label=i)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('时间')plt.ylabel('净值')plt.title('各组净值曲线')plt.show()<Figure size 432x288 with 0 Axes>
# 多空收益净值d = cum_df['G5'] - cum_df['G1']plt.figure()fig = plt.figure(figsize=(13, 5))ax1 = fig.add_subplot(1, 1, 1)ax1.plot(d, label='多组合')ax1.axhline(1.0, linestyle='-', color='black', lw=1)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('time')plt.ylabel('net value ratio')plt.title('多空净值组合')plt.show()<Figure size 432x288 with 0 Axes>
与在沪深300中的情况相同第四组年化收益及夏普较好
# 回测报告pd.DataFrame(report).T
.dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; }
| 夏普比率 | 年化收益率 | 日胜率 | 最大回撤 | 满仓天数 | 空仓天数 | 累计收益率 | |
|---|---|---|---|---|---|---|---|
| G1 | 0.11 | 0.35% | 50.26% | 35.05% | 764 | 620 | 2.03% |
| G2 | 0.57 | 9.63% | 50.51% | 41.51% | 891 | 493 | 69.52% |
| G3 | 0.31 | 3.86% | 53.63% | 33.67% | 647 | 737 | 24.32% |
| G4 | 0.98 | 16.61% | 51.03% | 26.46% | 925 | 459 | 141.61% |
| G5 | 0.75 | 11.32% | 51.60% | 34.39% | 843 | 541 | 85.07% |
通过敏感性分析我们可以看到在持有天数为25天时,夏普为**1.02**年化收益率为**19.32%**达到最优
# 夏普敏感性分析sharpe_df = threshold_analysis(factor_sz50, threshold_col='r_factor', params='sharpe')# 收益率敏感性分析ret_df = threshold_analysis(factor_sz50,threshold_col='r_factor',params='ret')print('夏普敏感性分析:')display(HTML(sharpe_df.to_html()))print('收率敏感性分析:')display(HTML(ret_df.to_html()))夏普敏感性分析:
| 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有25天 | |
|---|---|---|---|---|---|
| 277.6 | 0.660304 | 0.511365 | 0.525894 | 0.533767 | 0.568654 |
| 554.2 | 0.715452 | 0.633551 | 0.679475 | 0.676850 | 0.708128 |
| 830.8 | 0.911218 | 0.762223 | 0.881747 | 0.962861 | 1.027565 |
| 1107.4 | 0.585904 | 0.738515 | 0.734261 | 0.745644 | 0.960380 |
收率敏感性分析:
| 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有25天 | |
|---|---|---|---|---|---|
| 277.6 | 13.04% | 9.71% | 10.19% | 10.43% | 11.38% |
| 554.2 | 12.72% | 11.44% | 13.07% | 13.22% | 14.11% |
| 830.8 | 14.12% | 12.12% | 14.90% | 17.16% | 19.32% |
| 1107.4 | 7.41% | 10.42% | 10.68% | 11.31% | 15.85% |
# 获取因子factor_szzs = get_factor(szzs_datas, '000001.XSHG', 22)# 分组group_szzs = get_group(factor_szzs, 5, 'r_factor')group_szzs['未来5日涨幅'] = group_szzs.close.pct_change(4).shift(-4)group_szzs['未来15日涨幅'] = group_szzs.close.pct_change(14).shift(-14)group_szzs['未来20日涨幅'] = group_szzs.close.pct_change(19).shift(-19)# 绘图plt.figure(1, figsize=(12, 6))plt.title('当日涨幅')group_plot(group_szzs, '当日涨幅')plt.figure(2, figsize=(12, 6))plt.title('未来5日涨幅')group_plot(group_szzs, '未来5日涨幅')plt.figure(3, figsize=(12, 6))plt.title('未来15日涨幅')group_plot(group_szzs, '未来15日涨幅')plt.figure(4, figsize=(12, 6))plt.title('未来20日涨幅')group_plot(group_szzs, '未来20日涨幅')# 获取分组回测数据cum_ret_dic, report = group_back_test(group_szzs, 'GROUP', 20)# 构建dfcum_df = pd.DataFrame(cum_ret_dic, index=factor_szzs.index[:-1])# 获取分组名称group_list = cum_df.columns# 画图plt.figure()fig = plt.figure(figsize=(18, 8))ax1 = fig.add_subplot(1, 1, 1)for i in group_list:ax1.plot(cum_df[i], label=i)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('时间')plt.ylabel('净值')plt.title('各组净值曲线')plt.show()<Figure size 432x288 with 0 Axes>
# 多空收益净值d = cum_df['G5'] - cum_df['G1']plt.figure()fig = plt.figure(figsize=(13, 5))ax1 = fig.add_subplot(1, 1, 1)ax1.plot(d, label='多组合')ax1.axhline(1.0, linestyle='-', color='black', lw=1)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('time')plt.ylabel('net value ratio')plt.title('多空净值组合')plt.show()<Figure size 432x288 with 0 Axes>
情况与中证500上的效果相近，效果不理想
# 回测报告pd.DataFrame(report).T
.dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; }
| 夏普比率 | 年化收益率 | 日胜率 | 最大回撤 | 满仓天数 | 空仓天数 | 累计收益率 | |
|---|---|---|---|---|---|---|---|
| G1 | -0.17 | -3.27% | 52.26% | 32.03% | 685 | 699 | -17.38% |
| G2 | -0.094 | -3.37% | 53.29% | 44.78% | 820 | 564 | -17.85% |
| G3 | -0.12 | -3.69% | 54.65% | 62.93% | 591 | 793 | -19.43% |
| G4 | 0.053 | -0.85% | 54.78% | 45.35% | 889 | 495 | -4.80% |
| G5 | 0.33 | 3.73% | 54.02% | 36.57% | 870 | 514 | 23.42% |
通过敏感性分析我们可以看到在持有天数为5天时,夏普为**0.42**年化收益率为**6.18%**达到最优
# 夏普敏感性分析sharpe_df = threshold_analysis(factor_szzs, threshold_col='r_factor', params='sharpe')# 收益率敏感性分析ret_df = threshold_analysis(factor_szzs,threshold_col='r_factor',params='ret')print('夏普敏感性分析:')display(HTML(sharpe_df.to_html()))print('收率敏感性分析:')display(HTML(ret_df.to_html()))夏普敏感性分析:
| 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有25天 | |
|---|---|---|---|---|---|
| 277.6 | 0.326645 | 0.334628 | 0.316967 | 0.345330 | 0.350177 |
| 554.2 | 0.420137 | 0.341633 | 0.301278 | 0.281204 | 0.345446 |
| 830.8 | -0.006813 | -0.232146 | -0.189633 | -0.085984 | 0.067842 |
| 1107.4 | 0.596326 | 0.456162 | 0.232410 | 0.330138 | 0.384381 |
收率敏感性分析:
| 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有25天 | |
|---|---|---|---|---|---|
| 277.6 | 4.88% | 5.11% | 4.71% | 5.40% | 5.52% |
| 554.2 | 6.18% | 4.97% | 4.21% | 3.81% | 5.32% |
| 830.8 | -1.27% | -5.81% | -5.50% | -3.79% | -0.79% |
| 1107.4 | 5.77% | 4.83% | 2.24% | 3.73% | 4.68% |
# 获取因子factor_cyb = get_factor(cyb_datas, '399006.XSHE', 22)# 分组group_cyb = get_group(factor_cyb, 5, 'r_factor')group_cyb['未来5日涨幅'] = group_cyb.close.pct_change(4).shift(-4)group_cyb['未来15日涨幅'] = group_cyb.close.pct_change(14).shift(-14)group_cyb['未来20日涨幅'] = group_cyb.close.pct_change(19).shift(-19)# 绘图plt.figure(1, figsize=(12, 6))plt.title('当日涨幅')group_plot(group_cyb, '当日涨幅')plt.figure(2, figsize=(12, 6))plt.title('未来5日涨幅')group_plot(group_cyb, '未来5日涨幅')plt.figure(3, figsize=(12, 6))plt.title('未来15日涨幅')group_plot(group_cyb, '未来15日涨幅')plt.figure(4, figsize=(12, 6))plt.title('未来20日涨幅')group_plot(group_cyb, '未来20日涨幅')# 获取分组回测数据cum_ret_dic, report = group_back_test(group_cyb, 'GROUP', 20)# 构建dfcum_df = pd.DataFrame(cum_ret_dic, index=factor_cyb.index[:-1])# 获取分组名称group_list = cum_df.columns# 画图plt.figure()fig = plt.figure(figsize=(18, 8))ax1 = fig.add_subplot(1, 1, 1)for i in group_list:ax1.plot(cum_df[i], label=i)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('时间')plt.ylabel('净值')plt.title('各组净值曲线')plt.show()<Figure size 432x288 with 0 Axes>
# 多空收益净值d = cum_df['G5'] - cum_df['G1']plt.figure()fig = plt.figure(figsize=(13, 5))ax1 = fig.add_subplot(1, 1, 1)ax1.plot(d, label='多组合')ax1.axhline(1.0, linestyle='-', color='black', lw=1)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('time')plt.ylabel('net value ratio')plt.title('多空净值组合')plt.show()<Figure size 432x288 with 0 Axes>
# 回测报告pd.DataFrame(report).T
.dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; }
| 夏普比率 | 年化收益率 | 日胜率 | 最大回撤 | 满仓天数 | 空仓天数 | 累计收益率 | |
|---|---|---|---|---|---|---|---|
| G1 | 0.42 | 6.17% | 49.66% | 30.99% | 745 | 639 | 41.00% |
| G2 | 0.057 | -1.93% | 49.38% | 48.97% | 887 | 497 | -10.59% |
| G3 | 0.071 | -1.63% | 50.83% | 62.54% | 663 | 721 | -9.00% |
| G4 | 0.056 | -2.03% | 49.81% | 67.92% | 805 | 579 | -11.10% |
| G5 | 0.12 | 0.47% | 49.36% | 47.41% | 778 | 606 | 2.73% |
通过敏感性分析我们可以看到在持有天数为15天时,夏普为**0.31**年化收益率为**4.92%**达到最优
# 夏普敏感性分析sharpe_df = threshold_analysis(factor_cyb, threshold_col='r_factor', params='sharpe')# 收益率敏感性分析ret_df = threshold_analysis(factor_cyb,threshold_col='r_factor',params='ret')print('夏普敏感性分析:')display(HTML(sharpe_df.to_html()))print('收率敏感性分析:')display(HTML(ret_df.to_html()))夏普敏感性分析:
| 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有25天 | |
|---|---|---|---|---|---|
| 277.6 | 0.252302 | 0.302834 | 0.310161 | 0.296136 | 0.248519 |
| 554.2 | 0.119716 | 0.299526 | 0.238660 | 0.238902 | 0.211113 |
| 830.8 | -0.015989 | -0.057498 | 0.051276 | 0.118246 | 0.180309 |
| 1107.4 | 0.137731 | 0.230182 | 0.284131 | 0.119185 | 0.216516 |
收率敏感性分析:
| 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有25天 | |
|---|---|---|---|---|---|
| 277.6 | 3.06% | 4.68% | 4.92% | 4.45% | 2.88% |
| 554.2 | -0.67% | 4.58% | 2.66% | 2.62% | 1.71% |
| 830.8 | -2.70% | -4.55% | -2.26% | -0.78% | 0.90% |
| 1107.4 | 0.95% | 2.46% | 3.52% | 0.47% | 2.35% |
# 获取因子factor_zbzs = get_factor(zbzs_datas, '399101.XSHE', 22)# 分组group_zbzs = get_group(factor_zbzs, 5, 'r_factor')group_zbzs['未来5日涨幅'] = group_zbzs.close.pct_change(4).shift(-4)group_zbzs['未来15日涨幅'] = group_zbzs.close.pct_change(14).shift(-14)group_zbzs['未来20日涨幅'] = group_zbzs.close.pct_change(19).shift(-19)# 绘图plt.figure(1, figsize=(12, 6))plt.title('当日涨幅')group_plot(group_zbzs, '当日涨幅')plt.figure(2, figsize=(12, 6))plt.title('未来5日涨幅')group_plot(group_zbzs, '未来5日涨幅')plt.figure(3, figsize=(12, 6))plt.title('未来15日涨幅')group_plot(group_zbzs, '未来15日涨幅')plt.figure(4, figsize=(12, 6))plt.title('未来20日涨幅')group_plot(group_zbzs, '未来20日涨幅')# 获取分组回测数据cum_ret_dic, report = group_back_test(group_zbzs, 'GROUP', 20)# 构建dfcum_df = pd.DataFrame(cum_ret_dic, index=factor_zbzs.index[:-1])# 获取分组名称group_list = cum_df.columns# 画图plt.figure()fig = plt.figure(figsize=(18, 8))ax1 = fig.add_subplot(1, 1, 1)for i in group_list:ax1.plot(cum_df[i], label=i)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('时间')plt.ylabel('净值')plt.title('各组净值曲线')plt.show()<Figure size 432x288 with 0 Axes>
# 多空收益净值d = cum_df['G5'] - cum_df['G1']plt.figure()fig = plt.figure(figsize=(13, 5))ax1 = fig.add_subplot(1, 1, 1)ax1.plot(d, label='多组合')ax1.axhline(1.0, linestyle='-', color='black', lw=1)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('time')plt.ylabel('net value ratio')plt.title('多空净值组合')plt.show()<Figure size 432x288 with 0 Axes>
分组的单调性较好，但稳定性不太理想
# 回测报告pd.DataFrame(report).T
.dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; }
| 夏普比率 | 年化收益率 | 日胜率 | 最大回撤 | 满仓天数 | 空仓天数 | 累计收益率 | |
|---|---|---|---|---|---|---|---|
| G1 | -0.073 | -2.38% | 50.95% | 41.64% | 736 | 648 | -12.91% |
| G2 | -0.015 | -2.46% | 50.75% | 36.44% | 796 | 588 | -13.32% |
| G3 | 0.22 | 2.39% | 55.84% | 58.04% | 668 | 716 | 14.54% |
| G4 | 0.39 | 5.94% | 53.81% | 41.79% | 866 | 518 | 39.29% |
| G5 | 0.68 | 10.58% | 53.96% | 40.80% | 858 | 526 | 78.11% |
通过敏感性分析我们可以看到在持有天数为25天时,夏普为**0.67**年化收益率为**10.79%**达到最优
# 夏普敏感性分析sharpe_df = threshold_analysis(factor_zbzs, threshold_col='r_factor', params='sharpe')# 收益率敏感性分析ret_df = threshold_analysis(factor_zbzs,threshold_col='r_factor',params='ret')print('夏普敏感性分析:')display(HTML(sharpe_df.to_html()))print('收率敏感性分析:')display(HTML(ret_df.to_html()))夏普敏感性分析:
| 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有25天 | |
|---|---|---|---|---|---|
| 277.6 | 0.234493 | 0.306628 | 0.356705 | 0.401716 | 0.352142 |
| 554.2 | 0.119749 | 0.088895 | 0.182591 | 0.326514 | 0.313701 |
| 830.8 | 0.532049 | 0.310342 | 0.378942 | 0.516337 | 0.425004 |
| 1107.4 | 0.218446 | 0.462764 | 0.448659 | 0.677465 | 0.665997 |
收率敏感性分析:
| 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有25天 | |
|---|---|---|---|---|---|
| 277.6 | 2.74% | 4.69% | 6.09% | 7.38% | 5.99% |
| 554.2 | 0.07% | -0.87% | 1.38% | 5.24% | 4.90% |
| 830.8 | 7.75% | 4.19% | 5.79% | 9.04% | 7.19% |
| 1107.4 | 1.95% | 5.80% | 6.11% | 10.57% | 10.79% |
试图通过等权重通过CSAD和指数收益率排名打分构建成类似因子一样的值进行回测，收益率单调性还不错
# 采用等权重法将改进的两个指标*一个综合指标def get_score(data):rank_data1=data[['r_csad']].rank(ascending=False) #将csad因子按降序排名，排名越大越好rank_data2=data[['index_Nret']].rank(ascending=True) #将index_ret升序排名，排名越大越好rank_data=pd.concat([rank_data1,rank_data2],axis=1)data['SCORE']=rank_data.mean(axis=1) #等权重计算得分形成综合得分,得分越大越好return data
# 分组回测test=get_score(factor_hs300)g_test= get_group(test, 5, 'SCORE')# 获取分组回测数据cum_ret_dic, report = group_back_test(g_test, 'GROUP',20)# 构建dfcum_df = pd.DataFrame(cum_ret_dic, index=factor_hs300.index[:-1])# 获取分组名称group_list = cum_df.columns# 画图plt.figure()fig = plt.figure(figsize=(18, 8))ax1 = fig.add_subplot(1, 1, 1)for i in group_list:ax1.plot(cum_df[i], label=i)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('时间')plt.ylabel('净值')plt.title('各组净值曲线')plt.show()<Figure size 432x288 with 0 Axes>
# 回测报告pd.DataFrame(report).T
.dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; }
| 夏普比率 | 年化收益率 | 日胜率 | 最大回撤 | 满仓天数 | 空仓天数 | 累计收益率 | |
|---|---|---|---|---|---|---|---|
| G1 | 0.79 | 11.22% | 53.37% | 34.80% | 830 | 554 | 84.16% |
| G2 | 0.47 | 7.52% | 51.88% | 38.87% | 1091 | 293 | 51.62% |
| G3 | 0.54 | 9.68% | 52.61% | 46.73% | 1205 | 179 | 70.00% |
| G4 | 0.27 | 3.56% | 52.30% | 48.72% | 1044 | 340 | 22.25% |
| G5 | 0.09 | -0.11% | 52.03% | 43.48% | 765 | 619 | -0.62% |
# 夏普敏感性分析sharpe_df = threshold_analysis(test, threshold_col='r_factor', params='sharpe')# 收益率敏感性分析ret_df = threshold_analysis(test,threshold_col='r_factor',params='ret')print('夏普敏感性分析:')display(HTML(sharpe_df.to_html()))print('收率敏感性分析:')display(HTML(ret_df.to_html()))夏普敏感性分析:
| 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有25天 | |
|---|---|---|---|---|---|
| 277.6 | 0.383144 | 0.472864 | 0.496967 | 0.489153 | 0.484300 |
| 554.2 | 0.255122 | 0.326271 | 0.411510 | 0.404458 | 0.424605 |
| 830.8 | 0.716314 | 0.911090 | 1.024472 | 0.918649 | 0.761617 |
| 1107.4 | -0.022069 | 0.323722 | 0.418821 | 0.464447 | 0.589634 |
收率敏感性分析:
| 持有5天 | 持有10天 | 持有15天 | 持有20天 | 持有25天 | |
|---|---|---|---|---|---|
| 277.6 | 6.35% | 8.67% | 9.33% | 9.14% | 9.03% |
| 554.2 | 3.15% | 4.88% | 6.99% | 6.93% | 7.48% |
| 830.8 | 9.61% | 14.28% | 17.58% | 16.22% | 14.00% |
| 1107.4 | -0.82% | 3.31% | 4.88% | 5.86% | 8.05% |
我们选取了申万一级行业指数，将其按行业属性划分为周期、消费、成 长、金融五大类，通过 CCK 模型判断各申万一级行业指数成分股间是 否存在羊群效应，买入/卖出各大类行业中出现羊群效应的申万一级行业 指数。
策略步骤：计算向前 22 日（包括当日）每天的成分股组合截面绝对离散度 CSAD，OLS估计CCK模型中$R_{m,t}^2$,认为当日该组合存在羊群效应，根据22日内指数平均收益率的正负区分羊群效应发生时的市场趋势为上涨/下跌，买入/卖出标的指数并持仓22 交易日，持有期不重复开仓。
ps：过滤指数中上市不足三月\st\当日停牌股票
| 行业大类 | 申万一级行业 |
|---|---|
| 周期 | 采掘I, 化工I, 钢铁I, 有色金属I, 建筑建材I, 机械设备I, 交运设备I, 信息设备I |
| 公用事业I, 交通运输I, 房地产I, 综合I, 建筑材料I, 建筑装饰I, 电气设备I,机械设备I | |
| 消费 | 农林牧渔I, 家用电器I, 食品饮料I, 纺织服装I, 轻工制造I, 医药生物I, 商业贸易I,休闲服务I, 汽车I |
| 成长 | 电子I, 信息服务I, 国防军工I, 计算机I, 传媒I, 通信I |
| 金融 | 金融服务I, 银行I, 非银金融I |
industries_group_dic= dict(zip(['801020', '801030', '801040', '801050', '801060', '801730','801070', '801090', '801180', '801160','801100','801170','801230','801710','801720','801890']+['801880', '801110','801130', '801120', '801150', '801210', '801140', '801010','801200']+['801080', '801770', '801760', '801750', '801740','801220']+['801780','801790', '801190'],\['周期'] * 16 + ['消费'] * 9 + ['成长'] * 6 + ['金融'] * 3))
# 获取各行业指数信息def industry_index_close(start, end):begin = get_trade_days(end_date=start, count=22)[0]# 获取申万一级行业指数代码industries_id = get_industries(name='sw_l1').index.tolist()# 防止超限limit_row = len(get_trade_days(start_date=start, end_date=end))if limit_row > 3000:print('大于查询限制')temp = []for industry_code in industries_id:q = query(finance.SW1_DAILY_PRICE.date, finance.SW1_DAILY_PRICE.code, finance.SW1_DAILY_PRICE.name, finance.SW1_DAILY_PRICE.close, finance.SW1_DAILY_PRICE.change_pct).filter( finance.SW1_DAILY_PRICE.code == industry_code, finance.SW1_DAILY_PRICE.date >= begin, finance.SW1_DAILY_PRICE.date <= end)df = finance.run_query(q)temp.append(df)datas = pd.concat(temp)# 储存数据pkl_file = open('industry_index_close.pkl', 'wb')pickle.dump(datas, pkl_file)print('以储存数据：industry_index_close.pkl')return datas# 获取每日申万一级行业成分股'''区间日期return dictdic：key tradeDate value (industry_code,code)'''def get_industry_dic(start, end):import itertools# 获取申万一级行业指数代码industries_id = get_industries(name='sw_l1').index.tolist()begin = get_trade_days(end_date=start, count=22)[0]# 获取交易日期trade_list = [x.strftime('%Y-%m-%d')for x in get_trade_days(start_date=begin, end_date=end)]industry_dic = {}for day in trade_list:temp = []print('success', day)for industriesId in industries_id:stocks = get_industry_stocks(industriesId, date=day)if len(stocks) > 0:temp.extend(list(itertools.product([industriesId], stocks)))industry_dic[day] = temp# 储存信息pkl_file = open('industry_dic.pkl', 'wb')pickle.dump(industry_dic, pkl_file)print('以储存字典信息：industry_dic.pkl')return industry_dic# 通过字典数据下载行业股票每日收盘价'''dic：key tradeDate value (industry_code,code)-return df|index-code|index-code|close|industry_code过滤当日未交易，上市不足三月股票'''def get_industry_daily(dic):temp = []for day in dic.keys():# 过滤多个行业所属股票df1 = pd.DataFrame(dic[day], columns=['industry_code', 'code'])df1.drop_duplicates('code', inplace=True)# 获取当日股票列表stocks = df1.code.values.tolist()# 转为字典df1.set_index('code', inplace=True)industry_dic = df1.to_dict('index')# 过滤上市不足三月股票stocks = filter_now_share(stocks, begin_date=day, n=3 * 30)# 过滤当日未交易股票stocks = filter_paused_stocks(stocks, begin_date=day)# 获取股票信息df = get_price(stocks, end_date=day, count=1, fields='close').to_frame()df.index.names = ['date', 'code']# 获取过滤后的股票idx = pd.IndexSlicedf = df.loc[idx[:, stocks], :]# 添加行业代码## 如果未查询到则为0df['industry_code'] = list(map(lambda x: industry_dic.get(x, '0')['industry_code'],df.index.get_level_values(1)))temp.append(df)print('success', day)# 合并数据industry_daily = pd.concat(temp)# 储存数据pkl_file = open('industry_daily.pkl', 'wb')pickle.dump(industry_daily, pkl_file)print('储存下载数据：industry_daily.pkl')return industry_daily# 设置基准指标def get_benchmark(index, start, end):benchmark = get_price(index, start_date=start, end_date=end, fields='close')ret = benchmark.close.pct_change()ret = ret.fillna(0)position=np.ones(len(ret))cum_ret = (1 + ret).cumprod()# 最大回撤max_nv = np.maximum.accumulate(cum_ret)mdd = -np.min(cum_ret / max_nv - 1)# 胜率winning_ratio = np.sum(np.where(ret > 0, 1, 0)) / np.sum(position)# 计算年化收益率annual_ret = cum_ret[-1]**(240 / (len(cum_ret) - 5)) - 1# 计算累计收益率cum_ret_rate = cum_ret[-1] - 1# 夏普sharpe = ret.mean() / ret.std() * np.sqrt(240)df = pd.DataFrame({'年化收益率': '{:.2%}'.format(annual_ret),'累计收益率': '{:.2%}'.format(cum_ret_rate),'胜率': '{:.2%}'.format(winning_ratio),'开仓天数': np.sum(position),'最大回撤': '{:.2%}'.format(mdd),'夏普': '{:.2}'.format(sharpe)},index=['benchmark'])cum_ret.name = 'benchmark'return df, cum_retstart = '2014-01-01'end = '2019-08-31'# 获取行业成分股数据industry_dic = get_industry_dic(start, end)# 获取行业指数数据industry_index_data = industry_index_close(start, end)# 获取成分股收盘数据industry_daily = get_industry_daily(industry_dic)
以储存数据：industry_close.pkl success 2013-12-02 success 2013-12-03 success 2013-12-04 success 2013-12-05 success 2013-12-06 success 2013-12-09 success 2013-12-10 success 2013-12-11 success 2013-12-12 success 2013-12-13 success 2013-12-16 success 2013-12-17 success 2013-12-18 success 2013-12-19 success 2013-12-20 success 2013-12-23 success 2013-12-24 success 2013-12-25 success 2013-12-26 success 2013-12-27 success 2013-12-30 success 2013-12-31 success 2014-01-02 success 2014-01-03 success 2014-01-06 success 2014-01-07 success 2014-01-08 success 2014-01-09 success 2014-01-10 success 2014-01-13 success 2014-01-14 success 2014-01-15 success 2014-01-16 success 2014-01-17 success 2014-01-20 success 2014-01-21 success 2014-01-22 success 2014-01-23 success 2014-01-24 success 2014-01-27 success 2014-01-28 success 2014-01-29 success 2014-01-30 success 2014-02-07 success 2014-02-10 success 2014-02-11 success 2014-02-12 success 2014-02-13 success 2014-02-14 success 2014-02-17 success 2014-02-18 success 2014-02-19 success 2014-02-20 success 2014-02-21 success 2014-02-24 success 2014-02-25 success 2014-02-26 success 2014-02-27 success 2014-02-28 success 2014-03-03 success 2014-03-04 success 2014-03-05 success 2014-03-06 success 2014-03-07 success 2014-03-10 success 2014-03-11 success 2014-03-12 success 2014-03-13 success 2014-03-14 success 2014-03-17 success 2014-03-18 success 2014-03-19 success 2014-03-20 success 2014-03-21 success 2014-03-24 success 2014-03-25 success 2014-03-26 success 2014-03-27 success 2014-03-28 success 2014-03-31 success 2014-04-01 success 2014-04-02 success 2014-04-03 success 2014-04-04 success 2014-04-08 success 2014-04-09 success 2014-04-10 success 2014-04-11 success 2014-04-14 success 2014-04-15 success 2014-04-16 success 2014-04-17 success 2014-04-18 success 2014-04-21 success 2014-04-22 success 2014-04-23 success 2014-04-24 success 2014-04-25 success 2014-04-28 success 2014-04-29 success 2014-04-30 success 2014-05-05 success 2014-05-06 success 2014-05-07 success 2014-05-08 success 2014-05-09 success 2014-05-12 success 2014-05-13 success 2014-05-14 success 2014-05-15 success 2014-05-16 success 2014-05-19 success 2014-05-20 success 2014-05-21 success 2014-05-22 success 2014-05-23 success 2014-05-26 success 2014-05-27 success 2014-05-28 success 2014-05-29 success 2014-05-30 success 2014-06-03 success 2014-06-04 success 2014-06-05 success 2014-06-06 success 2014-06-09 success 2014-06-10 success 2014-06-11 success 2014-06-12 success 2014-06-13 success 2014-06-16 success 2014-06-17 success 2014-06-18 success 2014-06-19 success 2014-06-20 success 2014-06-23 success 2014-06-24 success 2014-06-25 success 2014-06-26 success 2014-06-27 success 2014-06-30 success 2014-07-01 success 2014-07-02 success 2014-07-03 success 2014-07-04 success 2014-07-07 success 2014-07-08 success 2014-07-09 success 2014-07-10 success 2014-07-11 success 2014-07-14 success 2014-07-15 success 2014-07-16 success 2014-07-17 success 2014-07-18 success 2014-07-21 success 2014-07-22 success 2014-07-23 success 2014-07-24 success 2014-07-25 success 2014-07-28 success 2014-07-29 success 2014-07-30 success 2014-07-31 success 2014-08-01 success 2014-08-04 success 2014-08-05 success 2014-08-06 success 2014-08-07 success 2014-08-08 success 2014-08-11 success 2014-08-12 success 2014-08-13 success 2014-08-14 success 2014-08-15 success 2014-08-18 success 2014-08-19 success 2014-08-20 success 2014-08-21 success 2014-08-22 success 2014-08-25 success 2014-08-26 success 2014-08-27 success 2014-08-28 success 2014-08-29 success 2014-09-01 success 2014-09-02 success 2014-09-03 success 2014-09-04 success 2014-09-05 success 2014-09-09 success 2014-09-10 success 2014-09-11 success 2014-09-12 success 2014-09-15 success 2014-09-16 success 2014-09-17 success 2014-09-18 success 2014-09-19 success 2014-09-22 success 2014-09-23 success 2014-09-24 success 2014-09-25 success 2014-09-26 success 2014-09-29 success 2014-09-30 success 2014-10-08 success 2014-10-09 success 2014-10-10 success 2014-10-13 success 2014-10-14 success 2014-10-15 success 2014-10-16 success 2014-10-17 success 2014-10-20 success 2014-10-21 success 2014-10-22 success 2014-10-23 success 2014-10-24 success 2014-10-27 success 2014-10-28 success 2014-10-29 success 2014-10-30 success 2014-10-31 success 2014-11-03 success 2014-11-04 success 2014-11-05 success 2014-11-06 success 2014-11-07 success 2014-11-10 success 2014-11-11 success 2014-11-12 success 2014-11-13 success 2014-11-14 success 2014-11-17 success 2014-11-18 success 2014-11-19 success 2014-11-20 success 2014-11-21 success 2014-11-24 success 2014-11-25 success 2014-11-26 success 2014-11-27 success 2014-11-28 success 2014-12-01 success 2014-12-02 success 2014-12-03 success 2014-12-04 success 2014-12-05 success 2014-12-08 success 2014-12-09 success 2014-12-10 success 2014-12-11 success 2014-12-12 success 2014-12-15 success 2014-12-16 success 2014-12-17 success 2014-12-18 success 2014-12-19 success 2014-12-22 success 2014-12-23 success 2014-12-24 success 2014-12-25 success 2014-12-26 success 2014-12-29 success 2014-12-30 success 2014-12-31 success 2015-01-05 success 2015-01-06 success 2015-01-07 success 2015-01-08 success 2015-01-09 success 2015-01-12 success 2015-01-13 success 2015-01-14 success 2015-01-15 success 2015-01-16 success 2015-01-19 success 2015-01-20 success 2015-01-21 success 2015-01-22 success 2015-01-23 success 2015-01-26 success 2015-01-27 success 2015-01-28 success 2015-01-29 success 2015-01-30 success 2015-02-02 success 2015-02-03 success 2015-02-04 success 2015-02-05 success 2015-02-06 success 2015-02-09 success 2015-02-10 success 2015-02-11 success 2015-02-12 success 2015-02-13 success 2015-02-16 success 2015-02-17 success 2015-02-25 success 2015-02-26 success 2015-02-27 success 2015-03-02 success 2015-03-03 success 2015-03-04 success 2015-03-05 success 2015-03-06 success 2015-03-09 success 2015-03-10 success 2015-03-11 success 2015-03-12 success 2015-03-13 success 2015-03-16 success 2015-03-17 success 2015-03-18 success 2015-03-19 success 2015-03-20 success 2015-03-23 success 2015-03-24 success 2015-03-25 success 2015-03-26 success 2015-03-27 success 2015-03-30 success 2015-03-31 success 2015-04-01 success 2015-04-02 success 2015-04-03 success 2015-04-07 success 2015-04-08 success 2015-04-09 success 2015-04-10 success 2015-04-13 success 2015-04-14 success 2015-04-15 success 2015-04-16 success 2015-04-17 success 2015-04-20 success 2015-04-21 success 2015-04-22 success 2015-04-23 success 2015-04-24 success 2015-04-27 success 2015-04-28 success 2015-04-29 success 2015-04-30 success 2015-05-04 success 2015-05-05 success 2015-05-06 success 2015-05-07 success 2015-05-08 success 2015-05-11 success 2015-05-12 success 2015-05-13 success 2015-05-14 success 2015-05-15 success 2015-05-18 success 2015-05-19 success 2015-05-20 success 2015-05-21 success 2015-05-22 success 2015-05-25 success 2015-05-26 success 2015-05-27 success 2015-05-28 success 2015-05-29 success 2015-06-01 success 2015-06-02 success 2015-06-03 success 2015-06-04 success 2015-06-05 success 2015-06-08 success 2015-06-09 success 2015-06-10 success 2015-06-11 success 2015-06-12 success 2015-06-15 success 2015-06-16 success 2015-06-17 success 2015-06-18 success 2015-06-19 success 2015-06-23 success 2015-06-24 success 2015-06-25 success 2015-06-26 success 2015-06-29 success 2015-06-30 success 2015-07-01 success 2015-07-02 success 2015-07-03 success 2015-07-06 success 2015-07-07 success 2015-07-08 success 2015-07-09 success 2015-07-10 success 2015-07-13 success 2015-07-14 success 2015-07-15 success 2015-07-16 success 2015-07-17 success 2015-07-20 success 2015-07-21 success 2015-07-22 success 2015-07-23 success 2015-07-24 success 2015-07-27 success 2015-07-28 success 2015-07-29 success 2015-07-30 success 2015-07-31 success 2015-08-03 success 2015-08-04 success 2015-08-05 success 2015-08-06 success 2015-08-07 success 2015-08-10 success 2015-08-11 success 2015-08-12 success 2015-08-13 success 2015-08-14 success 2015-08-17 success 2015-08-18 success 2015-08-19 success 2015-08-20 success 2015-08-21 success 2015-08-24 success 2015-08-25 success 2015-08-26 success 2015-08-27 success 2015-08-28 success 2015-08-31 success 2015-09-01 success 2015-09-02 success 2015-09-07 success 2015-09-08 success 2015-09-09 success 2015-09-10 success 2015-09-11 success 2015-09-14 success 2015-09-15 success 2015-09-16 success 2015-09-17 success 2015-09-18 success 2015-09-21 success 2015-09-22 success 2015-09-23 success 2015-09-24 success 2015-09-25 success 2015-09-28 success 2015-09-29 success 2015-09-30 success 2015-10-08 success 2015-10-09 success 2015-10-12 success 2015-10-13 success 2015-10-14 success 2015-10-15 success 2015-10-16 success 2015-10-19 success 2015-10-20 success 2015-10-21 success 2015-10-22 success 2015-10-23 success 2015-10-26 success 2015-10-27 success 2015-10-28 success 2015-10-29 success 2015-10-30 success 2015-11-02 success 2015-11-03 success 2015-11-04 success 2015-11-05 success 2015-11-06 success 2015-11-09 success 2015-11-10 success 2015-11-11 success 2015-11-12 success 2015-11-13 success 2015-11-16 success 2015-11-17 success 2015-11-18 success 2015-11-19 success 2015-11-20 success 2015-11-23 success 2015-11-24 success 2015-11-25 success 2015-11-26 success 2015-11-27 success 2015-11-30 success 2015-12-01 success 2015-12-02 success 2015-12-03 success 2015-12-04 success 2015-12-07 success 2015-12-08 success 2015-12-09 success 2015-12-10 success 2015-12-11 success 2015-12-14 success 2015-12-15 success 2015-12-16 success 2015-12-17 success 2015-12-18 success 2015-12-21 success 2015-12-22 success 2015-12-23 success 2015-12-24 success 2015-12-25 success 2015-12-28 success 2015-12-29 success 2015-12-30 success 2015-12-31 success 2016-01-04 success 2016-01-05 success 2016-01-06 success 2016-01-07 success 2016-01-08 success 2016-01-11 success 2016-01-12 success 2016-01-13 success 2016-01-14 success 2016-01-15 success 2016-01-18 success 2016-01-19 success 2016-01-20 success 2016-01-21 success 2016-01-22 success 2016-01-25 success 2016-01-26 success 2016-01-27 success 2016-01-28 success 2016-01-29 success 2016-02-01 success 2016-02-02 success 2016-02-03 success 2016-02-04 success 2016-02-05 success 2016-02-15 success 2016-02-16 success 2016-02-17 success 2016-02-18 success 2016-02-19 success 2016-02-22 success 2016-02-23 success 2016-02-24 success 2016-02-25 success 2016-02-26 success 2016-02-29 success 2016-03-01 success 2016-03-02 success 2016-03-03 success 2016-03-04 success 2016-03-07 success 2016-03-08 success 2016-03-09 success 2016-03-10 success 2016-03-11 success 2016-03-14 success 2016-03-15 success 2016-03-16 success 2016-03-17 success 2016-03-18 success 2016-03-21 success 2016-03-22 success 2016-03-23 success 2016-03-24 success 2016-03-25 success 2016-03-28 success 2016-03-29 success 2016-03-30 success 2016-03-31 success 2016-04-01 success 2016-04-05 success 2016-04-06 success 2016-04-07 success 2016-04-08 success 2016-04-11 success 2016-04-12 success 2016-04-13 success 2016-04-14 success 2016-04-15 success 2016-04-18 success 2016-04-19 success 2016-04-20 success 2016-04-21 success 2016-04-22 success 2016-04-25 success 2016-04-26 success 2016-04-27 success 2016-04-28 success 2016-04-29 success 2016-05-03 success 2016-05-04 success 2016-05-05 success 2016-05-06 success 2016-05-09 success 2016-05-10 success 2016-05-11 success 2016-05-12 success 2016-05-13 success 2016-05-16 success 2016-05-17 success 2016-05-18 success 2016-05-19 success 2016-05-20 success 2016-05-23 success 2016-05-24 success 2016-05-25 success 2016-05-26 success 2016-05-27 success 2016-05-30 success 2016-05-31 success 2016-06-01 success 2016-06-02 success 2016-06-03 success 2016-06-06 success 2016-06-07 success 2016-06-08 success 2016-06-13 success 2016-06-14 success 2016-06-15 success 2016-06-16 success 2016-06-17 success 2016-06-20 success 2016-06-21 success 2016-06-22 success 2016-06-23 success 2016-06-24 success 2016-06-27 success 2016-06-28 success 2016-06-29 success 2016-06-30 success 2016-07-01 success 2016-07-04 success 2016-07-05 success 2016-07-06 success 2016-07-07 success 2016-07-08 success 2016-07-11 success 2016-07-12 success 2016-07-13 success 2016-07-14 success 2016-07-15 success 2016-07-18 success 2016-07-19 success 2016-07-20 success 2016-07-21 success 2016-07-22 success 2016-07-25 success 2016-07-26 success 2016-07-27 success 2016-07-28 success 2016-07-29 success 2016-08-01 success 2016-08-02 success 2016-08-03 success 2016-08-04 success 2016-08-05 success 2016-08-08 success 2016-08-09 success 2016-08-10 success 2016-08-11 success 2016-08-12 success 2016-08-15 success 2016-08-16 success 2016-08-17 success 2016-08-18 success 2016-08-19 success 2016-08-22 success 2016-08-23 success 2016-08-24 success 2016-08-25 success 2016-08-26 success 2016-08-29 success 2016-08-30 success 2016-08-31 success 2016-09-01 success 2016-09-02 success 2016-09-05 success 2016-09-06 success 2016-09-07 success 2016-09-08 success 2016-09-09 success 2016-09-12 success 2016-09-13 success 2016-09-14 success 2016-09-19 success 2016-09-20 success 2016-09-21 success 2016-09-22 success 2016-09-23 success 2016-09-26 success 2016-09-27 success 2016-09-28 success 2016-09-29 success 2016-09-30 success 2016-10-10 success 2016-10-11 success 2016-10-12 success 2016-10-13 success 2016-10-14 success 2016-10-17 success 2016-10-18 success 2016-10-19 success 2016-10-20 success 2016-10-21 success 2016-10-24 success 2016-10-25 success 2016-10-26 success 2016-10-27 success 2016-10-28 success 2016-10-31 success 2016-11-01 success 2016-11-02 success 2016-11-03 success 2016-11-04 success 2016-11-07 success 2016-11-08 success 2016-11-09 success 2016-11-10 success 2016-11-11 success 2016-11-14 success 2016-11-15 success 2016-11-16 success 2016-11-17 success 2016-11-18 success 2016-11-21 success 2016-11-22 success 2016-11-23 success 2016-11-24 success 2016-11-25 success 2016-11-28 success 2016-11-29 success 2016-11-30 success 2016-12-01 success 2016-12-02 success 2016-12-05 success 2016-12-06 success 2016-12-07 success 2016-12-08 success 2016-12-09 success 2016-12-12 success 2016-12-13 success 2016-12-14 success 2016-12-15 success 2016-12-16 success 2016-12-19 success 2016-12-20 success 2016-12-21 success 2016-12-22 success 2016-12-23 success 2016-12-26 success 2016-12-27 success 2016-12-28 success 2016-12-29 success 2016-12-30 success 2017-01-03 success 2017-01-04 success 2017-01-05 success 2017-01-06 success 2017-01-09 success 2017-01-10 success 2017-01-11 success 2017-01-12 success 2017-01-13 success 2017-01-16 success 2017-01-17 success 2017-01-18 success 2017-01-19 success 2017-01-20 success 2017-01-23 success 2017-01-24 success 2017-01-25 success 2017-01-26 success 2017-02-03 success 2017-02-06 success 2017-02-07 success 2017-02-08 success 2017-02-09 success 2017-02-10 success 2017-02-13 success 2017-02-14 success 2017-02-15 success 2017-02-16 success 2017-02-17 success 2017-02-20 success 2017-02-21 success 2017-02-22 success 2017-02-23 success 2017-02-24 success 2017-02-27 success 2017-02-28 success 2017-03-01 success 2017-03-02 success 2017-03-03 success 2017-03-06 success 2017-03-07 success 2017-03-08 success 2017-03-09 success 2017-03-10 success 2017-03-13 success 2017-03-14 success 2017-03-15 success 2017-03-16 success 2017-03-17 success 2017-03-20 success 2017-03-21 success 2017-03-22 success 2017-03-23 success 2017-03-24 success 2017-03-27 success 2017-03-28 success 2017-03-29 success 2017-03-30 success 2017-03-31 success 2017-04-05 success 2017-04-06 success 2017-04-07 success 2017-04-10 success 2017-04-11 success 2017-04-12 success 2017-04-13 success 2017-04-14 success 2017-04-17 success 2017-04-18 success 2017-04-19 success 2017-04-20 success 2017-04-21 success 2017-04-24 success 2017-04-25 success 2017-04-26 success 2017-04-27 success 2017-04-28 success 2017-05-02 success 2017-05-03 success 2017-05-04 success 2017-05-05 success 2017-05-08 success 2017-05-09 success 2017-05-10 success 2017-05-11 success 2017-05-12 success 2017-05-15 success 2017-05-16 success 2017-05-17 success 2017-05-18 success 2017-05-19 success 2017-05-22 success 2017-05-23 success 2017-05-24 success 2017-05-25 success 2017-05-26 success 2017-05-31 success 2017-06-01 success 2017-06-02 success 2017-06-05 success 2017-06-06 success 2017-06-07 success 2017-06-08 success 2017-06-09 success 2017-06-12 success 2017-06-13 success 2017-06-14 success 2017-06-15 success 2017-06-16 success 2017-06-19 success 2017-06-20 success 2017-06-21 success 2017-06-22 success 2017-06-23 success 2017-06-26 success 2017-06-27 success 2017-06-28 success 2017-06-29 success 2017-06-30 success 2017-07-03 success 2017-07-04 success 2017-07-05 success 2017-07-06 success 2017-07-07 success 2017-07-10 success 2017-07-11 success 2017-07-12 success 2017-07-13 success 2017-07-14 success 2017-07-17 success 2017-07-18 success 2017-07-19 success 2017-07-20 success 2017-07-21 success 2017-07-24 success 2017-07-25 success 2017-07-26 success 2017-07-27 success 2017-07-28 success 2017-07-31 success 2017-08-01 success 2017-08-02 success 2017-08-03 success 2017-08-04 success 2017-08-07 success 2017-08-08 success 2017-08-09 success 2017-08-10 success 2017-08-11 success 2017-08-14 success 2017-08-15 success 2017-08-16 success 2017-08-17 success 2017-08-18 success 2017-08-21 success 2017-08-22 success 2017-08-23 success 2017-08-24 success 2017-08-25 success 2017-08-28 success 2017-08-29 success 2017-08-30 success 2017-08-31 success 2017-09-01 success 2017-09-04 success 2017-09-05 success 2017-09-06 success 2017-09-07 success 2017-09-08 success 2017-09-11 success 2017-09-12 success 2017-09-13 success 2017-09-14 success 2017-09-15 success 2017-09-18 success 2017-09-19 success 2017-09-20 success 2017-09-21 success 2017-09-22 success 2017-09-25 success 2017-09-26 success 2017-09-27 success 2017-09-28 success 2017-09-29 success 2017-10-09 success 2017-10-10 success 2017-10-11 success 2017-10-12 success 2017-10-13 success 2017-10-16 success 2017-10-17 success 2017-10-18 success 2017-10-19 success 2017-10-20 success 2017-10-23 success 2017-10-24 success 2017-10-25 success 2017-10-26 success 2017-10-27 success 2017-10-30 success 2017-10-31 success 2017-11-01 success 2017-11-02 success 2017-11-03 success 2017-11-06 success 2017-11-07 success 2017-11-08 success 2017-11-09 success 2017-11-10 success 2017-11-13 success 2017-11-14 success 2017-11-15 success 2017-11-16 success 2017-11-17 success 2017-11-20 success 2017-11-21 success 2017-11-22 success 2017-11-23 success 2017-11-24 success 2017-11-27 success 2017-11-28 success 2017-11-29 success 2017-11-30 success 2017-12-01 success 2017-12-04 success 2017-12-05 success 2017-12-06 success 2017-12-07 success 2017-12-08 success 2017-12-11 success 2017-12-12 success 2017-12-13 success 2017-12-14 success 2017-12-15 success 2017-12-18 success 2017-12-19 success 2017-12-20 success 2017-12-21 success 2017-12-22 success 2017-12-25 success 2017-12-26 success 2017-12-27 success 2017-12-28 success 2017-12-29 success 2018-01-02 success 2018-01-03 success 2018-01-04 success 2018-01-05 success 2018-01-08 success 2018-01-09 success 2018-01-10 success 2018-01-11 success 2018-01-12 success 2018-01-15 success 2018-01-16 success 2018-01-17 success 2018-01-18 success 2018-01-19 success 2018-01-22 success 2018-01-23 success 2018-01-24 success 2018-01-25 success 2018-01-26 success 2018-01-29 success 2018-01-30 success 2018-01-31 success 2018-02-01 success 2018-02-02 success 2018-02-05 success 2018-02-06 success 2018-02-07 success 2018-02-08 success 2018-02-09 success 2018-02-12 success 2018-02-13 success 2018-02-14 success 2018-02-22 success 2018-02-23 success 2018-02-26 success 2018-02-27 success 2018-02-28 success 2018-03-01 success 2018-03-02 success 2018-03-05 success 2018-03-06 success 2018-03-07 success 2018-03-08 success 2018-03-09 success 2018-03-12 success 2018-03-13 success 2018-03-14 success 2018-03-15 success 2018-03-16 success 2018-03-19 success 2018-03-20 success 2018-03-21 success 2018-03-22 success 2018-03-23 success 2018-03-26 success 2018-03-27 success 2018-03-28 success 2018-03-29 success 2018-03-30 success 2018-04-02 success 2018-04-03 success 2018-04-04 success 2018-04-09 success 2018-04-10 success 2018-04-11 success 2018-04-12 success 2018-04-13 success 2018-04-16 success 2018-04-17 success 2018-04-18 success 2018-04-19 success 2018-04-20 success 2018-04-23 success 2018-04-24 success 2018-04-25 success 2018-04-26 success 2018-04-27 success 2018-05-02 success 2018-05-03 success 2018-05-04 success 2018-05-07 success 2018-05-08 success 2018-05-09 success 2018-05-10 success 2018-05-11 success 2018-05-14 success 2018-05-15 success 2018-05-16 success 2018-05-17 success 2018-05-18 success 2018-05-21 success 2018-05-22 success 2018-05-23 success 2018-05-24 success 2018-05-25 success 2018-05-28 success 2018-05-29 success 2018-05-30 success 2018-05-31 success 2018-06-01 success 2018-06-04 success 2018-06-05 success 2018-06-06 success 2018-06-07 success 2018-06-08 success 2018-06-11 success 2018-06-12 success 2018-06-13 success 2018-06-14 success 2018-06-15 success 2018-06-19 success 2018-06-20 success 2018-06-21 success 2018-06-22 success 2018-06-25 success 2018-06-26 success 2018-06-27 success 2018-06-28 success 2018-06-29 success 2018-07-02 success 2018-07-03 success 2018-07-04 success 2018-07-05 success 2018-07-06 success 2018-07-09 success 2018-07-10 success 2018-07-11 success 2018-07-12 success 2018-07-13 success 2018-07-16 success 2018-07-17 success 2018-07-18 success 2018-07-19 success 2018-07-20 success 2018-07-23 success 2018-07-24 success 2018-07-25 success 2018-07-26 success 2018-07-27 success 2018-07-30 success 2018-07-31 success 2018-08-01 success 2018-08-02 success 2018-08-03 success 2018-08-06 success 2018-08-07 success 2018-08-08 success 2018-08-09 success 2018-08-10 success 2018-08-13 success 2018-08-14 success 2018-08-15 success 2018-08-16 success 2018-08-17 success 2018-08-20 success 2018-08-21 success 2018-08-22 success 2018-08-23 success 2018-08-24 success 2018-08-27 success 2018-08-28 success 2018-08-29 success 2018-08-30 success 2018-08-31 success 2018-09-03 success 2018-09-04 success 2018-09-05 success 2018-09-06 success 2018-09-07 success 2018-09-10 success 2018-09-11 success 2018-09-12 success 2018-09-13 success 2018-09-14 success 2018-09-17 success 2018-09-18 success 2018-09-19 success 2018-09-20 success 2018-09-21 success 2018-09-25 success 2018-09-26 success 2018-09-27 success 2018-09-28 success 2018-10-08 success 2018-10-09 success 2018-10-10 success 2018-10-11 success 2018-10-12 success 2018-10-15 success 2018-10-16 success 2018-10-17 success 2018-10-18 success 2018-10-19 success 2018-10-22 success 2018-10-23 success 2018-10-24 success 2018-10-25 success 2018-10-26 success 2018-10-29 success 2018-10-30 success 2018-10-31 success 2018-11-01 success 2018-11-02 success 2018-11-05 success 2018-11-06 success 2018-11-07 success 2018-11-08 success 2018-11-09 success 2018-11-12 success 2018-11-13 success 2018-11-14 success 2018-11-15 success 2018-11-16 success 2018-11-19 success 2018-11-20 success 2018-11-21 success 2018-11-22 success 2018-11-23 success 2018-11-26 success 2018-11-27 success 2018-11-28 success 2018-11-29 success 2018-11-30 success 2018-12-03 success 2018-12-04 success 2018-12-05 success 2018-12-06 success 2018-12-07 success 2018-12-10 success 2018-12-11 success 2018-12-12 success 2018-12-13 success 2018-12-14 success 2018-12-17 success 2018-12-18 success 2018-12-19 success 2018-12-20 success 2018-12-21 success 2018-12-24 success 2018-12-25 success 2018-12-26 success 2018-12-27 success 2018-12-28 success 2019-01-02 success 2019-01-03 success 2019-01-04 success 2019-01-07 success 2019-01-08 success 2019-01-09 success 2019-01-10 success 2019-01-11 success 2019-01-14 success 2019-01-15 success 2019-01-16 success 2019-01-17 success 2019-01-18 success 2019-01-21 success 2019-01-22 success 2019-01-23 success 2019-01-24 success 2019-01-25 success 2019-01-28 success 2019-01-29 success 2019-01-30 success 2019-01-31 success 2019-02-01 success 2019-02-11 success 2019-02-12 success 2019-02-13 success 2019-02-14 success 2019-02-15 success 2019-02-18 success 2019-02-19 success 2019-02-20 success 2019-02-21 success 2019-02-22 success 2019-02-25 success 2019-02-26 success 2019-02-27 success 2019-02-28 success 2019-03-01 success 2019-03-04 success 2019-03-05 success 2019-03-06 success 2019-03-07 success 2019-03-08 success 2019-03-11 success 2019-03-12 success 2019-03-13 success 2019-03-14 success 2019-03-15 success 2019-03-18 success 2019-03-19 success 2019-03-20 success 2019-03-21 success 2019-03-22 success 2019-03-25 success 2019-03-26 success 2019-03-27 success 2019-03-28 success 2019-03-29 success 2019-04-01 success 2019-04-02 success 2019-04-03 success 2019-04-04 success 2019-04-08 success 2019-04-09 success 2019-04-10 success 2019-04-11 success 2019-04-12 success 2019-04-15 success 2019-04-16 success 2019-04-17 success 2019-04-18 success 2019-04-19 success 2019-04-22 success 2019-04-23 success 2019-04-24 success 2019-04-25 success 2019-04-26 success 2019-04-29 success 2019-04-30 success 2019-05-06 success 2019-05-07 success 2019-05-08 success 2019-05-09 success 2019-05-10 success 2019-05-13 success 2019-05-14 success 2019-05-15 success 2019-05-16 success 2019-05-17 success 2019-05-20 success 2019-05-21 success 2019-05-22 success 2019-05-23 success 2019-05-24 success 2019-05-27 success 2019-05-28 success 2019-05-29 success 2019-05-30 success 2019-05-31 success 2019-06-03 success 2019-06-04 success 2019-06-05 success 2019-06-06 success 2019-06-10 success 2019-06-11 success 2019-06-12 success 2019-06-13 success 2019-06-14 success 2019-06-17 success 2019-06-18 success 2019-06-19 success 2019-06-20 success 2019-06-21 success 2019-06-24 success 2019-06-25 success 2019-06-26 success 2019-06-27 success 2019-06-28 success 2019-07-01 success 2019-07-02 success 2019-07-03 success 2019-07-04 success 2019-07-05 success 2019-07-08 success 2019-07-09 success 2019-07-10 success 2019-07-11 success 2019-07-12 success 2019-07-15 success 2019-07-16 success 2019-07-17 success 2019-07-18 success 2019-07-19 success 2019-07-22 success 2019-07-23 success 2019-07-24 success 2019-07-25 success 2019-07-26 success 2019-07-29 success 2019-07-30 success 2019-07-31 success 2019-08-01 success 2019-08-02 success 2019-08-05 success 2019-08-06 success 2019-08-07 success 2019-08-08 success 2019-08-09 success 2019-08-12 success 2019-08-13 success 2019-08-14 success 2019-08-15 success 2019-08-16 success 2019-08-19 success 2019-08-20 success 2019-08-21 success 2019-08-22 success 2019-08-23 success 2019-08-26 success 2019-08-27 success 2019-08-28 success 2019-08-29 success 2019-08-30 储存下载数据：industry_daily1.pkl
# 读取行业成分股相关数据pkl_file = open('industry_daily.pkl', 'rb')industry_daily = pickle.load(pkl_file)# 读取行业指数pkl_file = open('industry_index_close.pkl', 'rb')industry_index_data = pickle.load(pkl_file)# ind convert strindustry_index_data.code = industry_index_data.code.astype(str)# 添加大类## 指数数据industry_index_data['category_'] = industry_index_data.code.apply(lambda x: industries_group_dic[x])## 成分股数据industry_daily['category_'] = industry_daily['industry_code'].apply(lambda x: industries_group_dic[x])
# 计算csadindustry_index_data['change_pct'] = industry_index_data.change_pct / 100 # change_pct单位为%industry_index_ret = industry_index_data.groupby(['date', 'category_' ]).agg({'change_pct': mean})# 计算每日收益率(耗时较多。。。)#industry_daily=industry_daily.reset_index()#industry_daily['pct_change']=industry_daily.groupby(['code','date'])['close'].transform(lambda x:x.pct_change())
# 计算日收益率## industry_daily multi indexret_data = industry_daily[['close']].unstack()ret_data = ret_data.pct_change()ret_data = ret_data.stack()ret_data.columns = ['ret']# 添加retindustry_daily['ret'] = ret_dataindustry_daily = industry_daily.reset_index()industry_daily['date'] = industry_daily['date'].apply(lambda x: x.strftime('%Y-%m-%d'))# 行业指数收益率index转换industry_index_ret = industry_index_ret.reset_index()combine_data = pd.merge(industry_daily,industry_index_ret,on=['date', 'category_'])combine_data = combine_data[combine_data.date > '2013-12-02']# 计算差值的绝对值combine_data['diff_'] = abs(combine_data.ret - combine_data.change_pct)# 计算csadcsad = combine_data.groupby(['date', 'category_']).agg({'diff_': mean})csad.columns = ['CSAD']industry_index_ret = industry_index_ret.set_index(['date', 'category_'])csad['index_ret'] = industry_index_retcsad['index_ret_2'] = csad['index_ret']**2# 调整index级别csad = csad.swaplevel('category_', 'date')# 计算回归系数def lstsq(df):x = df[['abs_index_ret', 'index_ret_2']]y = df[['CSAD']]return np.linalg.lstsq(x, y)[0][1]singal_df = []idx = pd.IndexSlicecategory_list = csad.index.levels[0]# 取得回归系数for i in category_list:regression_df = csad.loc[idx[i, :]]# 回归时 指数收益为绝对值regression_df['abs_index_ret'] = abs(regression_df['index_ret'])temp = []# 滚动计算for r in range(len(regression_df) - 21):slice_data = regression_df[r:r + 21]temp.extend(lstsq(slice_data))# 添加回归系数regression_df['beta'] = np.r_[np.zeros(21), np.array(temp)]# 添加大类regression_df['categroy_'] = i# 添加先后regression_df['singal'] = (regression_df.beta < 0) & (regression_df.index_ret > 0)# 计算22日平均收益率regression_df['rolling_ret']=regression_df['index_ret'].rolling(21).mean()# 滚动信号regression_df['rolling_singal']=(regression_df.beta < 0) & (regression_df.rolling_ret > 0)# 根据指数22日平均收益开仓singal_df.append(regression_df)industry_singal_df = pd.concat(singal_df)
市场趋势为上涨时策略在行业指数上的平均表现如下所示（分年度表现 见附录）。市场趋势方面，市场存在上涨趋势时，羊群效应策略效果在周期行业较为明显。市值风格方面，高市值行业回测结果弱于低市值行业，金融型行业平均收益率为 10.20%，胜率 49.74%。行业属性方面，成长型行业表现较好，平均收益率18.29%，胜率 54.79%
# 根据指数22日平均收益开仓# 储存回测风险指标report_df = pd.DataFrame()# 储存净值用于作图net_dic = {}for k, group_df in industry_singal_df.groupby('categroy_'):group_df.index = group_df.index.astype('datetime64')ret = group_df['index_ret'].shift(-1)position = group_df['rolling_singal'] * np.ones(len(group_df['rolling_singal']))x_ret = np.nan_to_num(ret * position)cum_ret = (1 + x_ret).cumprod()# 最大回撤max_nv = np.maximum.accumulate(cum_ret)mdd = -np.min(cum_ret / max_nv - 1)# 胜率winning_ratio = np.sum(np.where(x_ret > 0, 1, 0)) / np.sum(position)# 计算年化收益率annual_ret = cum_ret[-1]**(240 / (len(cum_ret) - 5)) - 1# 计算累计收益率cum_ret_rate = cum_ret[-1] - 1# 夏普sharpe = x_ret.mean() / x_ret.std() * np.sqrt(240)df = pd.DataFrame({'年化收益率': '{:.2%}'.format(annual_ret),'累计收益率': '{:.2%}'.format(cum_ret_rate),'胜率': '{:.2%}'.format(winning_ratio),'开仓天数': np.sum(position),'最大回撤': '{:.2%}'.format(mdd),'夏普': '{:.2}'.format(sharpe)},index=[k])report_df = report_df.append(df)# 储存净值曲线net_dic[k] = cum_ret# 设置基准add_r, add_nev = get_benchmark('000300.XSHG', min(industry_singal_df.index), max(industry_singal_df.index))report_df = pd.concat([report_df, add_r]) # 添加基准# 查看风险指标print('各指数风险指标:')display(HTML(report_df.to_html()))nev_df = pd.DataFrame(net_dic, index=group_df.index)nev_df = nev_df.join(add_nev) # 添加基准# 画图plt.figure()fig = plt.figure(figsize=(18, 8))ax1 = fig.add_subplot(1, 1, 1)for i in nev_df.columns:ax1.plot(nev_df[i], label=i)ax1.xaxis.set_major_formatter(mdate.DateFormatter('%Y-%m'))plt.legend(loc='best')plt.xlabel('时间')plt.ylabel('净值')plt.title('各组净值曲线')plt.show()各指数风险指标:
| 年化收益率 | 累计收益率 | 胜率 | 开仓天数 | 最大回撤 | 夏普 | |
|---|---|---|---|---|---|---|
| 周期 | 15.51% | 131.79% | 57.27% | 784.0 | 29.55% | 0.93 |
| 成长 | 18.29% | 166.22% | 54.79% | 752.0 | 32.02% | 0.94 |
| 消费 | 16.62% | 145.02% | 57.28% | 817.0 | 25.30% | 1.0 |
| 金融 | 10.20% | 76.12% | 49.74% | 774.0 | 28.19% | 0.6 |
| benchmark | 7.87% | 55.54% | 52.42% | 1404.0 | 46.70% | 0.44 |
<Figure size 432x288 with 0 Axes>
本社区仅针对特定人员开放
查看需注册登录并通过风险意识测评
5秒后跳转登录页面...
