量化研究之成长股因子选股——模板化可定制
序
下面是博主之前在米匡量化平台做的一个因子研究，本来是自己写着玩玩，想来可能会有程序员朋友对此感兴趣的，就拿来分享一下。
平台有策略、研究两大块，策略部分分享一个博主的策略，现已上英雄榜可以订阅了。写这篇博客的时候打分榜第三名，让我嘚瑟一下。
由于平台经常更新，论坛的许多教程代码跑不通，博主还是以官方API文档为准，做了以下的选股模板。
亲测可以直接粘贴使用。选股板块、成长因子、时间均可自由发挥。
使用环境
米匡平台-研究板块
成长股因子选股
下面内容直接从研究板块导出md文件来的，排版较烂，结合注释理解吧，保留了部分结果显示。
#此处写一个成长股因子打分选股部分,考虑TMT、医疗器械板块，C34 通用设备制造业
C35 专用设备制造业
C36 汽车制造业
C37 铁路、船舶、航空航天和其它运输设备制造业
C38 电气机械及器材制造业
C39 计算机、通信和其他电子设备制造业
I63 电信、广播电视和卫星传输服务
I64 互联网和相关服务
I65 软件和信息技术服务业
#J66 货币金融服务
#J67 资本市场服务（由于选出的股票为银行类，这里未考虑J66/67，若是互联网金融方向可以考虑加入）
1.研发费用
2.市值大小
3.波动率（去掉，波动率分析在博主另一篇研究中）
4.利润率
5.pe
6.季度净利润同比增长率（Gr_Q_Earning）（单季度营业利润同比增长率（Gr_Q_OpEarning）或单季度营业收入同比增长率（Gr_Q_Sale））（未实现）
import talib
import pandas as pd
import numpy as np
import time
import math
import itertools
import matplotlib.pyplot as plt
stock_all=[]
industry_sector=['C34','C35','C36','C37','C38','C39','I63','I64','I65']
for i in industry_sector:
stock=industry(i)
for u in stock:
stock_all.append(u)
fundamental_df = get_fundamentals(
query(fundamentals.eod_derivative_indicator.market_cap,
fundamentals.eod_derivative_indicator.pe_ratio,
fundamentals.income_statement.r_n_d,
fundamentals.financial_indicator.net_profit_to_revenue
).filter(
fundamentals.income_statement.stockcode.in_(stock_all)
).filter(
fundamentals.eod_derivative_indicator.pe_ratio<100
).filter(
fundamentals.financial_indicator.net_profit_to_revenue>0
).filter(
fundamentals.eod_derivative_indicator.pe_ratio>0
). filter(
fundamentals.eod_derivative_indicator.market_cap>20000000000
).filter(
fundamentals.eod_derivative_indicator.market_cap<80000000000
).order_by(
fundamentals.eod_derivative_indicator.pe_ratio
),entry_date='2019-4-30',interval='1q'
,expect_df=False)
fundamental_df.dropna()
#npr净利润/营业总收入,rnd研发费用,pe动态市盈率
df_pe = fundamental_df['pe_ratio']
df_rnd = fundamental_df['r_n_d']
df_npr = fundamental_df['net_profit_to_revenue']
df_pe=df_pe.T
df_npr=df_npr.T
df_rnd=df_rnd.T
#修改列名字
df_pe.columns=['pe_ratio']
df_rnd.columns=['r_n_d']
df_npr.columns=['net_profit_to_revenue']
#将因子筛选的股票排序打分
def change_score(df):
#新建df_newdyr，排序，打分
df_newdyr=df.sort_values(by=df.columns[0],axis = 0,ascending = True)
df_newdyr['new_score']=range(1,len(df_newdyr)+1)
score_dyr=[]#新建股息率分数列，按原股票顺序排好，等会插入到df_dyr中
#遍历df_newdyr，获取每只股票的股息率分数
for i in range(0,len(df_newdyr)):
score_dyr.append(df_newdyr.loc[df.index[i]][1])
score_dyr=score_dyr
col_name=df.columns[0]+'_score'
df[col_name]=score_dyr
change_score(df_npr)
change_score(df_rnd)
df_pe['score_pe'] = range(len(df_pe),0,-1) #增加一列score
#合并4个因子打分表到df_score
df_score=df_pe.join(df_npr)
df_score=df_score.join(df_rnd)
df_score['sum_score']=df_score['score_pe']+df_score['net_profit_to_revenue_score']+df_score['r_n_d_score']
df_score=df_score.sort_values(by='sum_score',ascending=False)
print(df_score)
pe_ratio score_pe net_profit_to_revenue \
002624.XSHE 20.2827 74 23.8204
600741.XSHG 10.6567 93 6.5587
000581.XSHE 9.0107 94 30.8232
000425.XSHE 13.9095 90 7.338
600066.XSHG 13.6614 92 6.4961
002195.XSHE 15.6 83 37.2389
002236.XSHE 19.4177 75 7.0245
000100.XSHE 13.833 91 3.3931
002558.XSHE 38.3668 36 42.5748
601877.XSHG 15.4906 84 9.5351
000157.XSHE 14.632 87 10.8845
600487.XSHG 15.2499 85 6.5195
000418.XSHE 18.6541 76 9.3535
600271.XSHG 23.386 71 12.6534
600637.XSHG 18.2015 78 23.0387
600570.XSHG 52.0446 18 65.6754
002202.XSHE 14.7916 86 4.8468
002555.XSHE 27.6106 62 14.951
600528.XSHG 17.0272 80 7.6473
600699.XSHG 16.1644 82 3.2768
300296.XSHE 17.5456 79 15.2299
000413.XSHE 16.3028 81 11.8094
601869.XSHG 21.3844 73 15.4048
600522.XSHG 14.1968 89 4.7687
300003.XSHE 31.6784 48 30.803
603160.XSHG 50.5612 19 33.8194
600703.XSHG 21.4514 72 35.925
000938.XSHE 31.048 53 5.0535
300136.XSHE 25.6494 67 22.5093
002008.XSHE 27.5432 63 7.4863
... ... ... ...
300773.XSHE 40.0761 33 12.4464
300383.XSHE 36.9025 41 11.5507
002353.XSHE 31.5145 52 11.4972
002916.XSHE 43.4549 29 8.6252
603515.XSHG 28.3965 60 5.1726
000977.XSHE 47.2224 23 0.8898
300529.XSHE 58.1142 15 43.8276
002384.XSHE 33.4564 45 4.4626
300308.XSHE 46.5002 26 11.363
002129.XSHE 40.7193 32 6.1235
300253.XSHE 70.7282 10 20.7637
300766.XSHE 93.7558 1 42.8886
002465.XSHE 49.2613 20 3.1105
002153.XSHE 70.811 9 10.7012
002602.XSHE 54.0131 17 9.0277
002013.XSHE 31.624 49 2.4935
603986.XSHG 66.5266 13 8.6706
600320.XSHG 47.1206 24 1.4574
601179.XSHG 42.8337 30 3.0229
300024.XSHE 57.3201 16 10.6734
002049.XSHE 68.6453 12 9.988
600150.XSHG 76.378 5 0.1429
603019.XSHG 79.4685 3 2.32
600760.XSHG 44.6061 27 3.9009
600959.XSHG 43.5013 28 4.821
600118.XSHG 64.5165 14 5.3741
600893.XSHG 48.5943 21 0.287
600038.XSHG 48.3561 22 3.1286
000768.XSHE 77.2604 4 0.5511
600733.XSHG 82.0031 2 0.9936
net_profit_to_revenue_score r_n_d r_n_d_score sum_score
002624.XSHE 82 4.56704e+08 88 244
600741.XSHG 35 1.28332e+09 94 222
000581.XSHE 85 9.317e+07 41 220
000425.XSHE 39 4.52528e+08 87 216
600066.XSHG 33 4.1695e+08 84 209
002195.XSHE 89 7.21105e+07 32 204
002236.XSHE 38 5.41327e+08 90 203
000100.XSHE 18 1.16205e+09 93 202
002558.XSHE 90 1.91154e+08 72 198
601877.XSHG 51 1.53096e+08 61 196
000157.XSHE 56 1.15699e+08 52 195
600487.XSHG 34 2.25941e+08 76 195
000418.XSHE 49 1.82898e+08 70 195
600271.XSHG 64 1.41382e+08 58 193
600637.XSHG 81 7.24097e+07 33 192
600570.XSHG 94 3.1214e+08 80 192
002202.XSHE 27 2.46322e+08 78 191
002555.XSHE 69 1.46393e+08 59 190
600528.XSHG 42 1.68004e+08 67 189
600699.XSHG 16 5.27403e+08 89 187
300296.XSHE 70 8.18656e+07 36 185
000413.XSHE 61 9.67752e+07 42 184
601869.XSHG 72 8.39481e+07 39 184
600522.XSHG 25 1.6476e+08 66 180
300003.XSHE 84 1.08688e+08 47 179
603160.XSHG 86 2.08864e+08 73 178
600703.XSHG 88 3.41314e+07 14 174
000938.XSHE 28 9.02578e+08 92 173
300136.XSHE 79 6.13005e+07 26 172
002008.XSHE 41 1.7195e+08 68 172
... ... ... ... ...
300773.XSHE 63 6.34803e+07 27 123
300383.XSHE 60 4.95481e+07 21 122
002353.XSHE 59 2.93545e+07 11 122
002916.XSHE 44 1.05463e+08 46 119
603515.XSHG 29 6.57662e+07 29 118
000977.XSHE 4 4.31916e+08 85 112
300529.XSHE 92 1.33319e+07 5 112
002384.XSHE 23 8.19112e+07 37 105
300308.XSHE 57 4.94327e+07 20 103
002129.XSHE 32 8.34382e+07 38 102
300253.XSHE 76 3.89526e+07 16 102
300766.XSHE 91 2.34471e+07 9 101
002465.XSHE 14 1.53747e+08 62 96
002153.XSHE 55 7.05788e+07 31 95
002602.XSHE 46 6.67843e+07 30 93
002013.XSHE 8 8.04932e+07 34 91
603986.XSHG 45 6.40823e+07 28 86
600320.XSHG 6 1.2165e+08 53 83
601179.XSHG 13 8.09212e+07 35 78
300024.XSHE 54 1.08595e+07 4 74
002049.XSHE 53 1.00663e+07 3 68
600150.XSHG 1 1.48375e+08 60 66
603019.XSHG 7 1.1154e+08 50 60
600760.XSHG 20 2.841e+07 10 57
600959.XSHG 26 3.68412e+06 2 56
600118.XSHG 30 2.13345e+07 7 51
600893.XSHG 2 5.24764e+07 23 46
600038.XSHG 15 1.80412e+07 6 43
000768.XSHE 3 3.34097e+07 13 20
600733.XSHG 5 3.59638e+06 1 8
[94 rows x 7 columns]
#选择前5支股票
stocks=df_score.index[0:5].values
#选择排名倒数5支股票，建立对照组
stocks_test=df_score.index[-6:-1].values
print(stocks)
print(stocks_test)
['002624.XSHE' '600741.XSHG' '000581.XSHE' '000425.XSHE' '600066.XSHG']
['600959.XSHG' '600118.XSHG' '600893.XSHG' '600038.XSHG' '000768.XSHE']
#假设平均持有以上5支股票三个月，对比基准为上证300指数、创业板指数
#获取股票及指数每日收盘价，计算期间收益率。
stocks_bench=np.array(['000300.XSHG','399006.XSHE'])
bench_close=get_price(stocks_bench, start_date='2019-04-30', end_date='2019-07-30', frequency='1d', fields='close')
stocks_close=get_price(stocks, start_date='2019-04-30', end_date='2019-07-30', frequency='1d', fields='close')
test_close=get_price(stocks_test, start_date='2019-04-30', end_date='2019-07-30', frequency='1d', fields='close')
#定义求收益率函数
def a_return(close):
return (close.iloc[-1,:]-close.iloc[0,:])/close.iloc[0,:]
stocks_return=a_return(stocks_close)
bench_return=a_return(bench_close)
test_return=a_return(test_close)
#计算平均收益率
stocks_avg=stocks_return.sum()*0.2
test_avg=test_return.sum()*0.2
#绘制柱状图对比
data=np.append(bench_return,[stocks_avg,test_avg])
print(data)
data_bar = pd.Series(data, index=list(['沪深300指数','创业板指数','打分前5名','对照组test']))
data_bar.plot(kind='bar', figsize=(15, 6), fontsize=12)
plt.title('2019-4-30至2019-7-30收益率情况')
[-0.02688372 -0.01096118 -0.00875974 -0.01963428]
#可以看出基本面打分组收益率明显较高，由于是普跌的行情，应该说整体抗跌水平更优吧。
617
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
