请　[注册]　或　[登录]　 | 返回主站

量化交易吧 /  数理科学 帖子：3364764 新帖：2

波动和估值-1

Peace发表于：5 月 10 日 03：18回复(1)

展示回测数据

新手教程—收益率计算¶

以下内容主要介绍收益率计算

1 收益率计算¶

#导入需要的程序包
import pandas as pd
import seaborn as sns

df = get_price(get_industry_stocks('A01'), fields=('close',))['close']
df.head()
000998.XSHE 002041.XSHE 002772.XSHE 300087.XSHE 300189.XSHE 300511.XSHE 600108.XSHG 600313.XSHG 600354.XSHG 600359.XSHG 600371.XSHG 600506.XSHG 600540.XSHG 600598.XSHG 601118.XSHG
2015-01-05 19.44 12.71 NaN 4.00 2.71 NaN 9.51 4.50 8.81 10.92 10.55 11.33 6.28 9.34 8.88
2015-01-06 19.84 12.92 NaN 4.13 2.78 NaN 9.75 4.62 8.75 11.08 10.72 11.62 6.39 9.65 9.15
2015-01-07 19.68 12.85 NaN 4.08 2.76 NaN 9.98 4.70 8.85 11.07 10.70 11.54 6.37 9.68 9.03
2015-01-08 20.22 13.03 NaN 4.18 2.77 NaN 9.77 4.74 8.89 11.12 10.76 11.81 6.39 9.70 8.85
2015-01-09 19.86 12.87 NaN 4.06 2.72 NaN 9.41 4.59 8.60 10.80 10.47 11.73 6.30 9.58 8.61 
rets = df/df.shift(1) - 1#shift起平移作用
rets.head()
000998.XSHE 002041.XSHE 002772.XSHE 300087.XSHE 300189.XSHE 300511.XSHE 600108.XSHG 600313.XSHG 600354.XSHG 600359.XSHG 600371.XSHG 600506.XSHG 600540.XSHG 600598.XSHG 601118.XSHG
2015-01-05 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
2015-01-06 0.020576 0.016522 NaN 0.032500 0.025830 NaN 0.025237 0.026667 -0.006810 0.014652 0.016114 0.025596 0.017516 0.033191 0.030405
2015-01-07 -0.008065 -0.005418 NaN -0.012107 -0.007194 NaN 0.023590 0.017316 0.011429 -0.000903 -0.001866 -0.006885 -0.003130 0.003109 -0.013115
2015-01-08 0.027439 0.014008 NaN 0.024510 0.003623 NaN -0.021042 0.008511 0.004520 0.004517 0.005607 0.023397 0.003140 0.002066 -0.019934
2015-01-09 -0.017804 -0.012279 NaN -0.028708 -0.018051 NaN -0.036847 -0.031646 -0.032621 -0.028777 -0.026952 -0.006774 -0.014085 -0.012371 -0.027119 
returns = df.pct_change().dropna()
returns.head()
000998.XSHE 002041.XSHE 002772.XSHE 300087.XSHE 300189.XSHE 300511.XSHE 600108.XSHG 600313.XSHG 600354.XSHG 600359.XSHG 600371.XSHG 600506.XSHG 600540.XSHG 600598.XSHG 601118.XSHG 
# pandas.ols在0.20已被移除
## 2 移动窗口回归
# 使用普通最小二乘法(OLS)拟合曲线，得到回归系数及各类参数
# y = returns['300087.XSHE']
# x = returns.ix[:, ['300189.XSHE']]
# model = pd.ols(y=y, x=x)
# model
# model = pd.ols(y=y, x=x, window=5)
# model.beta.info()
# model.beta['300189.XSHE'].plot()
# 画出移动平均线、指数平滑移动平均线进行分析

df = get_price(get_industry_stocks('A01'), fields=('close',))['close']
plt.figure(figsize=[18,5])
df['000998.XSHE'].plot()
pd.rolling_mean(df['000998.XSHE'],20).plot(label='20 day moving average')
pd.rolling_mean(df['000998.XSHE'],5).plot(label='5 day moving average')
plt.legend(loc='best')

<matplotlib.legend.Legend at 0x7f20e9e4bad0>
df = get_price(get_industry_stocks('A01'), fields=('close',))['close']
plt.figure(figsize=[18,5])
df['000998.XSHE'].plot()
pd.rolling_mean(df['000998.XSHE'],20).plot(label='5 day moving average')
pd.ewma(df['000998.XSHE'],5).plot(label='5 day exponential moving average')
plt.legend(loc='best')

<matplotlib.legend.Legend at 0x7f20e9a59a90>
df = get_price(get_industry_stocks('A01'), fields=('close',))['close']
plt.figure(figsize=[18,5])
df['000998.XSHE'].plot()
pd.ewma(df['000998.XSHE'],20).plot(label='20 day exponential moving average')
pd.ewma(df['000998.XSHE'],5).plot(label='5 day exponential moving average')
plt.legend(loc='best')

<matplotlib.legend.Legend at 0x7f20e9a7a2d0>

       

全部回复

0/140

量化课程

    移动端课程