预测 statsmodel 参数错误 [英] predict statsmodel argument Error

问题描述

我正在尝试预测数组的样本外值.Python代码:

I am trying to predict outofsample values for an array. Python code:

import pandas as pd
import numpy as np
from statsmodels.tsa.arima_model import ARIMA

    dates = pd.date_range('2012-07-09','2012-07-30')
    series = [43.,32.,63.,98.,65.,78.,23.,35.,78.,56.,45.,45.,56.,6.,63.,45.,64.,34.,76.,34.,14.,54.]
    res = pd.Series(series, index=dates)
    r = ARIMA(res,(1,2,0))
    pred = r.predict(start='2012-07-31', end='2012-08-31')

我收到此错误.我看到我给出了两个参数，但编译器返回我给出了 3 个.

I am getting this error.I see I have given two argument but compiler return I have given 3.

Traceback (most recent call last):
  File "XXXXXXXXX/testfile.py", line 12, in <module>
    pred = r.predict(start='2012-07-31', end='2012-08-31')
TypeError: predict() takes at least 2 arguments (3 given)

请帮忙

推荐答案

ARIMA.predict 的调用签名是

predict(self, params, start=None, end=None, exog=None, dynamic=False)

因此，当您调用 r.predict(start='2012-07-31', end='2012-08-31') 时，self 被绑定到 r，并且值绑定到 start 和 end 但所需的位置参数 params 没有绑定.这就是为什么你得到错误

Thus, when you call r.predict(start='2012-07-31', end='2012-08-31'), self gets bound to r, and values are bound to start and end but the required positional arument params does not get bound. That is why you get the error

TypeError: predict() takes at least 2 arguments (3 given)

不幸的是，错误消息具有误导性.3给定"指的是r、start和end.2 个参数"是指两个必需的参数，self 和 params.问题是没有给出必需的位置参数params.

Unfortunately the error message is misleading. The "3 given" refer to r, start and end. The "2 arguments" refer to the two required arguments, self and params. The problem is that the required positional argument params was not given.

要解决问题，您需要参数.通常你通过拟合找到这些参数:

To fix the problem, you need parameters. Usually you find those parameters by fitting:

r = r.fit()

在调用之前

pred = r.predict(start='2012-07-31', end='2012-08-31')

r.fit() 返回一个 statsmodels.tsa.arima_model.ARIMAResultsWrapper有参数烘焙"所以调用 ARIMAResultWrapper.fit 不需要传递 params.

r.fit() returns a statsmodels.tsa.arima_model.ARIMAResultsWrapper which have the parameters "baked in" so calling ARIMAResultWrapper.fit does not require passing params.

import pandas as pd
import numpy as np
from statsmodels.tsa.arima_model import ARIMA

dates = pd.date_range('2012-07-09','2012-07-30')
series = [43.,32.,63.,98.,65.,78.,23.,35.,78.,56.,45.,45.,56.,6.,63.,45.,64.,34.,76.,34.,14.,54.]
res = pd.Series(series, index=dates)
r = ARIMA(res,(1,2,0))
r = r.fit()
pred = r.predict(start='2012-07-31', end='2012-08-31')
print(pred)

收益

2012-07-31   -39.067222
2012-08-01    26.902571
2012-08-02   -17.027333
...
2012-08-29     0.532946
2012-08-30     0.532447
2012-08-31     0.532780
Freq: D, dtype: float64

这篇关于预测 statsmodel 参数错误的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！