同花顺Supermind量化交易 运用SVM算法“打板”-- SVM算法应用

课程简介：介绍股票涨跌停板机制，对股票涨停进行分类，并研究股票涨停后的收益情况，最终采用SVM算法进行“打板”研究。

dt = dataclosedf
dt = dt.sort_values(by='date')

import matplotlib.pyplot as plt
import pandas as pd
import numpy as np


plt.style.use('seaborn')
fig = plt.figure()
axes = fig.add_axes([0.1, 0.1, 1, 0.618]) #插入面板

x1_list=list(dt['1日收盘价'])
y=np.array(x1_list)
x=np.array(range(0,len(x1_list)))
axes.scatter(x,y,c='tomato')
axes.set_xlabel('time',fontsize=15)
axes.set_ylabel('down_up',fontsize=15)
axes.set_title(' one day distribution',fontsize=18)
# #设置X轴
axes.set_xticks([0,100,200,300,400,500])
axes.set_xticklabels([list(dt.index)[0],list(dt.index)[100],list(dt.index)[200],list(dt.index)[300],list(dt.index)[400],list(dt.index)[500]],fontsize=15)

plt.style.use('seaborn')
fig = plt.figure()
axes = fig.add_axes([0.1, 0.1, 1, 0.618]) #插入面板

x1_list=list(dt['2日收盘价'])
y=np.array(x1_list)
x=np.array(range(0,len(x1_list)))
axes.scatter(x,y,c='tomato')
axes.set_xlabel('time',fontsize=15)
axes.set_ylabel('down_up',fontsize=15)
axes.set_title(' two day distribution',fontsize=18)
# #设置X轴
axes.set_xticks([0,100,200,300,400,500])
axes.set_xticklabels([list(dt.index)[0],list(dt.index)[100],list(dt.index)[200],list(dt.index)[300],list(dt.index)[400],list(dt.index)[500]],fontsize=15)

plt.style.use('seaborn')
fig = plt.figure()
axes = fig.add_axes([0.1, 0.1, 1, 0.618]) #插入面板

x1_list=list(dt['3日收盘价'])
y=np.array(x1_list)
x=np.array(range(0,len(x1_list)))
axes.scatter(x,y,c='tomato')
axes.set_xlabel('time',fontsize=15)
axes.set_ylabel('down_up',fontsize=15)
axes.set_title(' three day distribution',fontsize=18)
# #设置X轴
axes.set_xticks([0,100,200,300,400,500])
axes.set_xticklabels([list(dt.index)[0],list(dt.index)[100],list(dt.index)[200],list(dt.index)[300],list(dt.index)[400],list(dt.index)[500]],fontsize=15)

Out[13]:

[<matplotlib.text.Text at 0x7f1b30cd7630>,
 <matplotlib.text.Text at 0x7f1b313e20b8>,
 <matplotlib.text.Text at 0x7f1b30b9b630>,
 <matplotlib.text.Text at 0x7f1b313b9cc0>,
 <matplotlib.text.Text at 0x7f1b31aee0f0>,
 <matplotlib.text.Text at 0x7f1b30c90b00>]

In [14]:

labeldt = dataclosedf
labeldt['1日收盘价'] = labeldt['1日收盘价'].apply(lambda x:1 if x>0 else -1)
labeldt['2日收盘价'] = labeldt['2日收盘价'].apply(lambda x:1 if x>0 else -1)
labeldt['3日收盘价'] = labeldt['3日收盘价'].apply(lambda x:1 if x>0 else -1)
labeldt = labeldt.sort_values(by='date')
labeldt

Out[14]:

	stock	date	1日收盘价	2日收盘价	3日收盘价
2014-01-03	600010.SH	2014-01-03	-1	-1	-1
2014-01-16	000767.SZ	2014-01-16	-1	-1	-1
2014-02-12	000971.SZ	2014-02-12	-1	-1	-1
2014-02-21	000554.SZ	2014-02-21	-1	-1	-1
2014-02-28	600680.SH	2014-02-28	1	1	1
2014-03-10	300116.SZ	2014-03-10	-1	-1	-1
2014-03-13	000923.SZ	2014-03-13	-1	1	1
2014-03-27	000687.SZ	2014-03-27	-1	-1	-1
2014-04-03	000918.SZ	2014-04-03	-1	-1	-1
2014-04-10	300278.SZ	2014-04-10	1	1	1
2014-04-30	601001.SH	2014-04-30	1	-1	-1
2014-05-21	002660.SZ	2014-05-21	1	1	1
2014-06-04	600601.SH	2014-06-04	-1	-1	-1
2014-06-05	600074.SH	2014-06-05	-1	-1	-1
2014-06-06	002535.SZ	2014-06-06	1	1	1
2014-06-16	300198.SZ	2014-06-16	-1	-1	-1
2014-06-30	002093.SZ	2014-06-30	-1	-1	-1
2014-07-02	300288.SZ	2014-07-02	1	1	1
2014-07-04	002070.SZ	2014-07-04	-1	-1	-1
2014-07-07	600568.SH	2014-07-07	-1	1	1
2014-07-08	600435.SH	2014-07-08	1	1	1
2014-07-11	002629.SZ	2014-07-11	1	-1	-1
2014-07-16	000868.SZ	2014-07-16	-1	-1	-1
2014-07-25	600319.SH	2014-07-25	-1	1	1
2014-07-28	600157.SH	2014-07-28	1	1	1
2014-08-07	600556.SH	2014-08-07	-1	-1	-1
2014-08-11	600222.SH	2014-08-11	-1	-1	-1
2014-08-14	300208.SZ	2014-08-14	-1	1	1
2014-08-18	002125.SZ	2014-08-18	1	-1	1
2014-08-21	000922.SZ	2014-08-21	1	1	1
...	...	...	...	...	...
2018-12-19	002845.SZ	2018-12-19	1	-1	1
2018-12-25	000531.SZ	2018-12-25	-1	-1	-1
2018-12-26	300687.SZ	2018-12-26	-1	-1	-1
2018-12-27	300407.SZ	2018-12-27	-1	-1	-1
2018-12-28	601619.SH	2018-12-28	-1	-1	-1
2019-01-04	300125.SZ	2019-01-04	1	1	1
2019-01-07	300588.SZ	2019-01-07	-1	-1	-1
2019-01-08	600677.SH	2019-01-08	-1	-1	-1
2019-01-09	600452.SH	2019-01-09	1	1	1
2019-01-10	300076.SZ	2019-01-10	-1	-1	-1
2019-01-11	002941.SZ	2019-01-11	-1	-1	-1
2019-01-14	603220.SH	2019-01-14	-1	-1	1
2019-01-16	600446.SH	2019-01-16	-1	-1	-1
2019-01-18	600470.SH	2019-01-18	1	1	1
2019-01-21	300693.SZ	2019-01-21	-1	1	1
2019-01-22	300503.SZ	2019-01-22	-1	-1	-1
2019-01-24	002170.SZ	2019-01-24	-1	-1	-1
2019-01-25	600721.SH	2019-01-25	1	-1	-1
2019-01-28	000637.SZ	2019-01-28	-1	-1	-1
2019-01-31	300250.SZ	2019-01-31	1	1	1
2019-02-01	002011.SZ	2019-02-01	-1	-1	-1
2019-02-11	600532.SH	2019-02-11	-1	-1	-1
2019-02-12	600318.SH	2019-02-12	1	-1	-1
2019-02-13	300097.SZ	2019-02-13	1	1	1
2019-02-14	300256.SZ	2019-02-14	-1	1	1
2019-02-18	002099.SZ	2019-02-18	1	1	-1
2019-02-19	300468.SZ	2019-02-19	-1	-1	-1
2019-02-20	600393.SH	2019-02-20	-1	-1	1
2019-02-21	002565.SZ	2019-02-21	1	1	-1
2019-02-22	300063.SZ	2019-02-22	1	-1	-1

503 rows × 5 columns

In [15]:

tradeday = list(get_trade_days(startdate, '20200202', count=None).strftime('%Y-%m-%d'))

timelist = []
trlist= []
lplist =[]
q_trlist = []
for d in tradeday:
    time = 0
    tr = 0
    lp = 0
    q_tr = 0
    if d in list(labeldt['date']):
        stock = labeldt['stock'][d]
        day = tradeday[tradeday.index(d)+1]
        stockdata = get_price(stock,None,day,'1m',['close','high','low','open','turnover_rate'