有数据统计高手吗 协整性里面的P怎么求 有python参考

独倚斜阳

2组数据 通涨同跌  计算相关性 几乎是相关性超级高  但是他们不具备协整性  协整性的P值非常大



同理 这蓝色和红色线  不同涨跌  几乎没有相关性  但是具有协整性

所谓的相关性 我用老外的公式会计算   但是这个协整性的P值  我翻了很久没找到我能理解的公式 基本都是python的函数  看不懂不会计算 大致相当于这种

  

import pandas as pd
import numpy as np
from scipy.stats import pearsonr
from statsmodels.tsa.stattools import coint
import matplotlib.pyplot as plt
x1 = pd.Series (np.random.normal (0,1,1000))
x2 = x1.copy ()
for i in range (10):
x2[ (100*i): (100*i+100)] = (-1)**i

plt.figure (figsize= (6,3))
plt.plot (x1, c= ' blue')
plt.plot (x2, c= ' red')
plt.show ()   
_, pv, _ = coint (x1, x2)
print ("Cointegration test p-value : ", pv)
correl, pv = pearsonr (x1, x2)
print ("Correlation : ", correl)

import pandas as pd<br /> import numpy as np<br /> from scipy.stats import pearsonr<br /> from statsmodels.tsa.stattools import coint <br /> import matplotlib.pyplot as plt<br /> <br /> x1 = pd.Series(np.random.normal(0,1,1000))<br /> x2 = x1.copy()<br /> for i in range(10):<br />     x2[(100*i):(100*i+100)] = (-1)**i<br />     <br /> plt.figure(figsize=(6,3))<br /> plt.plot(x1, c='blue')<br /> plt.plot(x2, c='red')<br /> plt.show()    <br /> <br /> _, pv, _ = coint(x1, x2)<br /> print("Cointegration test p-value : ", pv)<br /> correl, pv = pearsonr(x1, x2)<br /> print("Correlation : ", correl)

也就是
_, pv, _ = coint(x1, x2)
print("Cointegration test p-value : ", pv)


有没有大佬能改写个易语言版本的  只要PV那个计算就行了，

子程序 传入2个文本型的参数数组 分别代表2组数据 返回协整性的P值

文章参考
https://zhuanlan.zhihu.com/p/158871695

明天自然醒

def coint(y0, y1, trend='c', method='aeg', maxlag=None, autolag='aic',
          return_results=None):
    """Test for no-cointegration of a univariate equation
    The null hypothesis is no cointegration. Variables in y0 and y1 are
    assumed to be integrated of order 1, I(1).
    This uses the augmented Engle-Granger two-step cointegration test.
    Constant or trend is included in 1st stage regression, i.e. in
    cointegrating equation.
    **Warning:** The autolag default has changed compared to statsmodels 0.8.
    In 0.8 autolag was always None, no the keyword is used and defaults to
    'aic'. Use `autolag=None` to avoid the lag search.
    Parameters
    ----------
    y1 : array_like, 1d
        first element in cointegrating vector
    y2 : array_like
        remaining elements in cointegrating vector
    trend : str {'c', 'ct'}
        trend term included in regression for cointegrating equation
        * 'c' : constant
        * 'ct' : constant and linear trend
        * also available quadratic trend 'ctt', and no constant 'nc'
    method : string
        currently only 'aeg' for augmented Engle-Granger test is available.
        default might change.
    maxlag : None or int
        keyword for `adfuller`, largest or given number of lags
    autolag : string
        keyword for `adfuller`, lag selection criterion.
        * if None, then maxlag lags are used without lag search
        * if 'AIC' (default) or 'BIC', then the number of lags is chosen
          to minimize the corresponding information criterion
        * 't-stat' based choice of maxlag.  Starts with maxlag and drops a
          lag until the t-statistic on the last lag length is significant
          using a 5%-sized test
    return_results : bool
        for future compatibility, currently only tuple available.
        If True, then a results instance is returned. Otherwise, a tuple
        with the test outcome is returned.
        Set `return_results=False` to avoid future changes in return.
    Returns
    -------
    coint_t : float
        t-statistic of unit-root test on residuals
    pvalue : float
        MacKinnon's approximate, asymptotic p-value based on MacKinnon (1994)
    crit_value : dict
        Critical values for the test statistic at the 1 %, 5 %, and 10 %
        levels based on regression curve. This depends on the number of
        observations.
    Notes
    -----

独倚斜阳

我在想 易语言这方面内容太少了 我要不要学一下调用python  这样用易语言调用 就好解决了 程序也不需要大改动

likeliang123

等待大佬解决。。

独倚斜阳

再问一下 如果要用易语言调用python的话 是不是运行exe的电脑 也需要单独装一遍python的环境？

明天自然醒

按原理一算不就好了吗

明天自然醒

    .. [1] MacKinnon, J.G. 1994  "Approximate Asymptotic Distribution Functions
       for Unit-Root and Cointegration Tests." Journal of Business & Economics
       Statistics, 12.2, 167-76.
    .. [2] MacKinnon, J.G. 2010.  "Critical Values for Cointegration Tests."
       Queen"s University, Dept of Economics Working Papers 1227.
       http://ideas.repec.org/p/qed/wpaper/1227.html

独倚斜阳

搞定了 学会了python了 然后就是调用问题了  模块那个  我装完python下了很多包后就没法正常初始化了 只有用精易支持库去