%matplotlib inline
import matplotlib.pyplot as plt
import seaborn
# plt.style.use('seaborn') ggplot;
import matplotlib as mpl
mpl.rcParams['font.family'] = 'serif' #解决一些字体显示乱码问题
import warnings; warnings.simplefilter('ignore')
import numpy as np
import pandas as pd
import tushare as ts
# Tushare API
前导知识学习
data = ts.get_k_data('600030', start = '2010-01-01', end='2017-06-30')
data.head() #DataFrame数据结构
data.set_index('date', inplace = True) #设置索引；替换，真实覆盖；
data['SMA_20'] = data['close'].rolling(20).mean()
data['SMA_60'] = data['close'].rolling(60).mean()
data[['close','SMA_20','SMA_60']].plot(figsize = (10,6)) #可视化
#计算股票return
data['returns'] = np.log(data['close'] / data['close'].shift(1))
#Numpy向量化操作，避免循环；
data['returns_dis'] = data['close']/data['close'].shift(1)-1
data['return_dis2'] = data['close'].pct_change()
data['position'] = np.where(data['SMA_20'] > data['SMA_60'], 1, -1)
data['returns'].cumsum().apply(np.exp).plot(figsize=(10, 6))
#可视化；计算累计收益，连续下的算法；
SMA策略
1. 数据准备 & 回测准备
import numpy as np
import pandas as pd
import tushare as ts
data = ts.get_k_data('hs300', start = '2010-01-01', end='2017-06-30')
#要用Tushare新的数据获取接口，不然数据获取有bug；
data = pd.DataFrame(data)
data.rename(columns={'close': 'price'}, inplace=True) #dict
data.set_index('date', inplace = True) #覆盖操作；
data.set_index('date')
data['SMA_10'] = data['price'].rolling(10).mean() #必背
data['SMA_60'] = data['price'].rolling(60).mean()
data[['price','SMA_10','SMA_60']].plot(title='HS300 stock price | 10 & 60 days SMAs',
figsize=(10, 6))
#Pandas里面如何选择多列必须要掌握；
2. 策略开发思路
data['position'] = np.where(data['SMA_10'] > data['SMA_60'], 1, -1)
data.dropna(inplace=True) #去掉空值，NaN
data['position'].plot(ylim=[-1.1, 1.1], title='Market Positioning')
3. 计算策略年化收益并可视化
data['returns'] = np.log(data['price'] / data['price'].shift(1))
#Numpy向量化；循环做法（尽量避免）,连续return
data['strategy'] = data['position'].shift(1) * data['returns']
#注意未来函数；一般会使得回测收益高估；
data[['returns', 'strategy']].sum()
data[['returns', 'strategy']].cumsum().apply(np.exp).plot(figsize=(10, 6)) #可视化；离散的计算方法参考Momoentum策略
4. 策略收益风险评估
data[['returns', 'strategy']].mean() * 252 #年化收益率；
data[['returns', 'strategy']].std() * 252 ** 0.5 #年化风险；
data['cumret'] = data['strategy'].cumsum().apply(np.exp)
data['cummax'] = data['cumret'].cummax()
data[['cumret', 'cummax']].plot(figsize=(10, 6))
drawdown = (data['cummax'] - data['cumret'])
drawdown.max() #计算原理：最大回撤；
5. 策略优化的一种思路
hs300 = ts.get_k_data('hs300','2010-01-01', '2017-06-30')[['date','close']]
# hs300 = pd.DataFrame(hs300)
hs300.rename(columns={'close': 'price'}, inplace=True)
hs300.set_index('date',inplace = True)
hs300.head()
hs300['SMA_10'] = hs300['price'].rolling(10).mean() #必背
hs300['SMA_60'] = hs300['price'].rolling(60).mean() #必背
hs300[['price', 'SMA_10', 'SMA_60']].tail()
hs300['10-60'] = hs300['SMA_10'] - hs300['SMA_60']
SD = 20 #阈值
hs300['regime'] = np.where(hs300['10-60'] > SD, 1,0)
hs300['regime'] = np.where(hs300['10-60'] < -SD, -1,hs300['regime']) #核心重要；
hs300['regime'].value_counts()
hs300['Market'] = np.log(hs300['price']/hs300['price'].shift(1))
hs300['Strategy'] = hs300['regime'].shift(1) * hs300['Market']
hs300[['Market','Strategy']].cumsum().apply(np.exp).plot(grid=True, figsize = (10,8))
