StatsModels 置信区间

未分类

Confidence Interval：（区间估计）
品种收益置信区间

import tushare as ts
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from scipy import stats
data = ts.get_k_data('000001', start='2016-01-01', end='2016-12-31', ktype='D',autype='qfq')
data.index = pd.to_datetime(data['date'],format='%Y-%m-%d')
per = data['close'].pct_change()[1:]
p_mu = np.mean(per)
p_std = np.std(per)
cil, cir = stats.norm.interval(0.95, loc=p_mu, scale=p_std)
print(cil, cir)

输出结果：说明用一年的数据计算出这只股票一天的涨跌幅95%的概率在（-2.27%，2.27%)之间

(-0.022749357724022465, 0.022726966268070206)

x = np.linspace(per.min(), per.max(), len(per))
y = stats.norm.pdf(x,p_mu, p_std)
plt.plot(x, y)
plt.vlines(cil,0, 35, colors='r', linestyles='dashed')
plt.vlines(cir, 0, 35, colors='r', linestyles='dashed')
plt.show()

Sample Mean vs. Population Mean

sample_size = 100
sample = np.random.normal(p_mu, p_std, size=sample_size)
s_mu = np.mean(sample)
s_std = np.std(sample)
se = s_std/np.sqrt(sample_size)
cil, cir = stats.t.interval(0.95, len(sample)-1, loc=s_mu, scale=se)
print(cil, cir)
x = np.linspace(sample.min(), sample.max(), len(sample)-1)
y = stats.norm.pdf(x, np.mean(sample), np.std(sample))
plt.plot(x, y)
plt.vlines(cil,0, 20, colors='r', linestyles='dashed')
plt.vlines(cir, 0, 20, colors='r', linestyles='dashed')
plt.show()

策略收益置信区间