1. import package and download data
from atrader import *import numpy as npimport pandas as pdimport matplotlib.pyplot as pltimport mathimport statsmodels.api as smimport datetime as dtimport scipy.stats as statsimport seaborn as sns# 获取因子数据# 单日多标的多因子# 4个月的动量类因子names = ['REVS60','REVS120','BIAS60','CCI20','PVT','MA10Close','DEA','RC20','RSTR63','DDI']factors1 = get_factor_by_day(factor_list= names, target_list=list(get_code_list('hs300',date='-02-01').code), date='-02-28')factors2 = get_factor_by_day(factor_list= names, target_list=list(get_code_list('hs300',date='-02-01').code), date='-03-29')factors3 = get_factor_by_day(factor_list= names, target_list=list(get_code_list('hs300',date='-02-01').code), date='-04-30')factors4 = get_factor_by_day(factor_list= names, target_list=list(get_code_list('hs300',date='-02-01').code), date='-05-31')
2. factors preprocess（maybe not）
# MAD:中位数去极值def extreme_MAD(dt,n = 5.2):median = dt.quantile(0.5) # 找出中位数new_median = (abs((dt - median)).quantile(0.5)) # 偏差值的中位数dt_up = median + n*new_median # 上限dt_down = median - n*new_median # 下限return dt.clip(dt_down, dt_up, axis=1) # 超出上下限的值，赋值为上下限# Z值标准化def standardize_z(dt):mean = dt.mean()# 截面数据均值std = dt.std() # 截面数据标准差return (dt - mean)/std# 行业中性化shenwan_industry = {'SWNLMY1':'sse.801010','SWCJ1':'sse.801020','SWHG1':'sse.801030','SWGT1':'sse.801040','SWYSJS1':'sse.801050','SWDZ1':'sse.801080','SWJYDQ1':'sse.801110','SWSPCL1':'sse.801120','SWFZFZ1':'sse.801130','SWQGZZ1':'sse.801140','SWYYSW1':'sse.801150','SWGYSY1':'sse.801160','SWJTYS1':'sse.801170','SWFDC1':'sse.801180','SWSYMY1':'sse.801200','SWXXFW1':'sse.801210','SWZH1':'sse.801230','SWJZCL1':'sse.801710','SWJZZS1':'sse.801720','SWDQSB1':'sse.801730','SWGFJG1':'sse.801740','SWJSJ1':'sse.801750','SWCM1':'sse.801760','SWTX1':'sse.801770','SWYH1':'sse.801780','SWFYJR1':'sse.801790','SWQC1':'sse.801880','SWJXSB1':'sse.801890'}# 构造行业哑变量矩阵def industry_exposure(target_idx):# 构建DataFrame，存储行业哑变量df = pd.DataFrame(index = [x.lower() for x in target_idx],columns = shenwan_industry.keys())for m in df.columns: # 遍历每个行业# 行标签集合和某个行业成分股集合的交集temp = list(set(df.index).intersection(set(get_code_list(m).code.tolist())))df.loc[temp, m] = 1# 将交集的股票在这个行业中赋值为1return df.fillna(0) # 将 NaN 赋值为0# 需要传入单个因子值和总市值def neutralization(factor,MktValue,industry = True):Y = factor.fillna(0)df = pd.DataFrame(index = Y.index, columns = Y.columns) # 构建输出矩阵for i in range(Y.shape[1]): # 遍历每一天的截面数据if type(MktValue) == pd.DataFrame:lnMktValue = MktValue.iloc[:,i].apply(lambda x:math.log(x)) # 市值对数化lnMktValue = lnMktValue.fillna(0)if industry: # 行业、市值dummy_industry = industry_exposure(Y.index.tolist())X = pd.concat([lnMktValue,dummy_industry],axis = 1,sort = False) # 市值与行业合并else: # 仅市值X = lnMktValueelif industry: # 仅行业dummy_industry = industry_exposure(factor.index.tolist())X = dummy_industry# X = sm.add_constant(X)result = sm.OLS(Y.iloc[:,i].astype(float),X.astype(float)).fit() # 线性回归df.iloc[:,i] = result.resid # 每日的截面数据存储到df中return df
3. corr（mean） test
# 计算因子的相关系数矩阵函数def factor_corr(factors):factors = factors.set_index('code')factors_process = standardize_z(extreme_MAD(factors.fillna(0)))result = factors_process.fillna(0).corr()return result# 获取相关系数矩阵factors_corr1 = factor_corr(factors1)factors_corr2 = factor_corr(factors2)factors_corr3 = factor_corr(factors3)factors_corr4 = factor_corr(factors4)factors_corr = (factors_corr1+factors_corr2+factors_corr3+factors_corr4).div(4) # 矩阵均值检验# 相关系数检验abs(factors_corr).mean()abs(factors_corr).median()
4. corr hot map plot
# 画图二fig = plt.figure()plt.subplots(figsize=(8, 6.4)) # 设置画面大小sns.heatmap(factors_corr, annot=True, vmax=1, vmin=-1, square=True, cmap="CMRmap_r",)plt.show()
5. OLS predict yeild between factors and stock return
# 因子合成corrnames = ['REVS60','REVS120','BIAS60','CCI20','MA10Close','DEA','RC20','DDI'] # 共线因子# 历史收益率加权法data = get_kdata_n(target_list=list(get_code_list('hs300').code), frequency='month', fre_num=1, n=5, end_date='-06-01', fill_up=False, df=True, fq=1)close = data.pivot_table(values='close',index='code',columns='time')# 数据透视，形成收盘价dataframecode = sorted(set(list(data['code'])),key =list(data['code']).index) # 数据透视形成的行标签排序与前面不一致，将原始数据的排序去重复项后排序不变stock_close = close.loc[code] # 形成行标签与前面一致的dataframestock_return = stock_close.diff(axis=1).div(stock_close) # 利用收盘价计算股票的月收益率#因子暴露与股票收益率回归估计收益率factors4 = factors4.set_index('code').rename(index = str.lower).loc[:,corrnames] # 提取共线的因子数据factor_return= list()for i in range(factors4.shape[1]):X = factors4.iloc[:,i]result = sm.OLS(stock_return.iloc[:,-1].astype(float),X.astype(float)).fit() # 股票收益率和因子数据回归factor_return.append(result.params[0])# 回归系数，即因子收益率
6. use yield combine factors
weight = [x/sum(map(abs,factor_return)) for x in factor_return] # 因子收益率的比重for i in range(factors4.shape[1]):factors4.iloc[:, i] = factors4.iloc[:,i].mul(abs(weight[i])) # 按因子收益率的占比计算各因子值composite_factorreturn = factors4.sum(axis=1) # 因子值合成print(composite_factorreturn)
