对比不同因子IC

1. 读取本地数据
2. 数据预处理
3. 计算月度IC
4. 对比不同因子的IC值
5. IC的时间序列
6. 计算各个因子的IR

1_读取本地数据

from jaqs.data.dataapi import DataApi
from jaqs.data import DataView
import numpy as np
from datetime import datetime
import pandas as pd
import warnings
import alphalens
warnings.filterwarnings("ignore")
dataview_folder = 'JAQS_Data/hs300'
dv = DataView()
dv.load_dataview(dataview_folder)

D:\Anaconda3\lib\site-packages\statsmodels\compat\pandas.py:56: FutureWarning: The pandas.core.datetools module is deprecated and will be removed in a future version. Please use the pandas.tseries module instead.
  from pandas.core import datetools


Dataview loaded successfully.

2_数据预处理

def change_columns_index(signal):
    new_names = {}
    for c in signal.columns:
        if c.endswith('SZ'):
            new_names[c] = c.replace('SZ', 'XSHE')
        elif c.endswith('SH'):
            new_names[c] = c.replace('SH', 'XSHG')
    signal = signal.rename_axis(new_names, axis=1)
    signal.index = pd.Index(map(lambda x: datetime.strptime(str(x),"%Y%m%d") , signal.index))
    return signal

all_factor = ['pb', 'roe', 'roa', 'price_div_dps', 'ps_ttm', 'pe_ttm']

origin_factors = {f: change_columns_index(dv.get_ts(f)) for f in all_factor}

from jaqs.research.signaldigger import process
#去极值,z_score标准化,加干扰值
PN_disturbed = pd.Panel({name: process.get_disturbed_factor(process.standardize(process.winsorize(frame)))\
                         for name, frame in origin_factors.items()})

print(PN_disturbed)

<class 'pandas.core.panel.Panel'>
Dimensions: 6 (items) x 1453 (major_axis) x 539 (minor_axis)
Items axis: pb to roe
Major_axis axis: 2012-01-04 00:00:00 to 2017-12-22 00:00:00
Minor_axis axis: 000001.XSHE to 603993.XSHG

prices = change_columns_index(dv.get_ts('close_adj'))

3_计算月度IC

alphalens.performance.mean_information_coefficient()

def cal_monthly_ic(factor_df):
    factor_data = alphalens.utils.get_clean_factor_and_forward_returns(factor_df.stack(), prices, quantiles=5)
    return alphalens.performance.mean_information_coefficient(factor_data, by_time='M')
monthly_ic = {key: cal_monthly_ic(value) for key, value in PN_disturbed.iteritems()}

print(list(monthly_ic.values())[0].tail())

                  1         5         10
date                                    
2017-08-31 -0.001214  0.062826  0.089109
2017-09-30  0.032383  0.094269  0.137698
2017-10-31  0.041499  0.111624  0.134414
2017-11-30 -0.040708 -0.087116 -0.116298
2017-12-31  0.094886  0.264356  0.234456

monthly_ic_mean = pd.DataFrame(
    list(map(lambda frame: frame.mean(), monthly_ic.values())),
    monthly_ic.keys()
)
monthly_ic_std = pd.DataFrame(
    list(map(lambda frame: frame.std(), monthly_ic.values())),
    monthly_ic.keys()
)

print(monthly_ic_mean)
print (monthly_ic_std)

                     1         5         10
pb            -0.013308 -0.010537 -0.014940
pe_ttm        -0.008865 -0.010849 -0.016026
price_div_dps -0.004332 -0.003527 -0.006607
ps_ttm        -0.004566 -0.002501 -0.004181
roa            0.008087  0.018150  0.023119
roe            0.006245  0.013959  0.020110
                     1         5         10
pb             0.055441  0.123580  0.159507
pe_ttm         0.037962  0.083811  0.112212
price_div_dps  0.029059  0.064610  0.086014
ps_ttm         0.044336  0.097744  0.127836
roa            0.034680  0.080584  0.105429
roe            0.030441  0.070708  0.097227

4_对比不同因子的IC值

import matplotlib.pyplot as plt
import numpy as np
fig, ax = plt.subplots(figsize=(15,7))
N=6
ind = np.arange(N)  # the x locations for the groups
width = 0.3       # the width of the bars
# Example data
ind_name = tuple(monthly_ic_mean.index)
y_pos = np.arange(len(ind))
one_mean = monthly_ic_mean.iloc[:,0]
one_std = monthly_ic_std.iloc[:,0]
five_mean = monthly_ic_mean.iloc[:,1]
five_std = monthly_ic_std.iloc[:,1]
ten_mean = monthly_ic_mean.iloc[:,2]
ten_std = monthly_ic_std.iloc[:,2]
ax.barh(ind - width, one_mean, align='edge',height=0.2, xerr=one_std, label='one_day', color='r')
ax.barh(ind - 0.05, five_mean, align='edge',height=0.2,xerr=five_std, label='five_day', color='y')
ax.barh(ind + width, ten_mean, align='center',height=0.2,xerr=ten_std, label='ten_day', color='b')
ax.set_yticks(y_pos)
ax.set_yticklabels(ind_name)
ax.invert_yaxis()  # labels read top-to-bottom
ax.set_xlabel('mean_IC')
ax.set_title('HS300_IC in Different Period')
plt.legend()
plt.show()

5_IC的时间序列

PN_IC = pd.Panel(monthly_ic)

fig = plt.figure(figsize=(15,7))
plt.plot(PN_IC.minor_xs(5))
plt.hlines(0.02,PN_IC.minor_xs(5).index[0],PN_IC.minor_xs(5).index[-1] , linestyles='dashed', alpha=0.5)
plt.legend(PN_IC.minor_xs(5).columns)
plt.title('FiveDay_IC')
plt.show()

6_计算各个因子的IR

def IR(df):
    return df.mean()/df.std()
print ('OneDay_IR:\n', IR(PN_IC.minor_xs(1)))
print ('FiveDay_IR:\n', IR(PN_IC.minor_xs(5)))
print ('TenDay_IR:\n', IR(PN_IC.minor_xs(10)))

OneDay_IR:
 pb              -0.240034
pe_ttm          -0.233529
price_div_dps   -0.149082
ps_ttm          -0.102978
roa              0.233188
roe              0.205164
dtype: float64
FiveDay_IR:
 pb              -0.085268
pe_ttm          -0.129441
price_div_dps   -0.054597
ps_ttm          -0.025585
roa              0.225227
roe              0.197414
dtype: float64
TenDay_IR:
 pb              -0.093661
pe_ttm          -0.142818
price_div_dps   -0.076814
ps_ttm          -0.032705
roa              0.219280
roe              0.206836
dtype: float64