Python 的季节性 ARIMA，频率 H 不明白？

Question

我是新来的。我正在尝试按照此 TIME SERIE DECOMPOSITION EXAMPLE with this CSV DATA.

来分解时间序列

我的问题出在从 statsmodels.tsa.seasonal 导入的 season_decompose 函数中。我试图弄清楚如何将它应用于我的数据但没有成功。这是我的代码：

import os
import csv
import time
import datetime
import pandas as pd
import numpy as np
import statsmodels.api as sm

from datetime import datetime
from datetime import timedelta, date
from dateutil.relativedelta import relativedelta
from statsmodels.tsa.seasonal import seasonal_decompose

import matplotlib.pyplot as plt
import matplotlib.dates as mdates

from itertools import product

df = pd.read_csv('table.csv', index_col=0)
df.index.name=None
df.reset_index(inplace=True)

start = datetime.strptime("2015-10-10", "%Y-%m-%d")
date_list = [start + relativedelta(days =x , hour=y) for x,y in product(range(0,93), range(0,24))]

df['index'] =date_list
df.set_index(['index'], inplace=True)
df.index.name=None
df.columns= ['Close']
df['Close'] = df.Close.apply(lambda x: int(x))
df.Close.plot(figsize=(12,8), title= 'Monthly Closehip', fontsize=14)


decomposition = seasonal_decompose(df.Close, freq=93)  

fig = decomposition.plot()  
fig.set_size_inches(15, 8)
plt.show()

我收到以下错误：

Traceback (most recent call last):
File "test.py", line 59, in <module>
    decomposition = seasonal_decompose(df.Close, freq=93)  
  File "/usr/local/lib/python2.7/dist-packages/statsmodels/tsa/seasonal.py", line 70, in seasonal_decompose
    pfreq = freq_to_period(pfreq)
  File "/usr/local/lib/python2.7/dist-packages/statsmodels/tsa/tsatools.py", line 657, in freq_to_period
    "think this in error.".format(freq))
ValueError: freq H not understood. Please report if you think this in error.

数据是一个 csv 文件：https://docs.google.com/a/esi.dz/spreadsheets/d/1s2Ak6Rqgm43FV4G_J_giWeHyi38xdZCBCz2v34k7iuA/edit?usp=sharing

请尝试帮助我。

Answer 1

在查看了一些博客并测试了一些解决方案之后。我带来了这个：

将值添加到df.Close允许像这样进行分解：

decomposition = seasonal_decompose(df.Close.values, freq=168)

Python 的季节性 ARIMA，频率 H 不明白？

Seasonal ARIMA with Python, freq H not understood?

python

statistics

time-series

statsmodels