基于ARIMAX的多变量预测模型python源码+数据集+代码注释.zip

import pandas as pd import numpy as np from statsmodels.tsa.arima_model import ARIMA from statsmodels.graphics.tsaplots import plot_acf,plot_pacf import matplotlib.pyplot as plt from matplotlib.pylab import style import statsmodels.api as sm from itertools import product import pyflux as pf plt.rcParams['font.sans-serif'] = ['Microsoft YaHei'] listlabel=['猪肉价格','仔猪价格','玉米市场价格','通货膨胀率CPI','鸡肉价格','存栏量'] def datacf(): df=pd.read_csv("data2.csv",encoding='gbk') df['时间'] = pd.to_datetime(df['时间']) df.index = df['时间'] df = df.resample('3M').mean() ## 1:单位根检验检验序列的平稳性,ADF 检验 for i in listlabel: df[i]=df[i].diff() df[i]=df[i].dropna() df["存栏量"]=df["存栏量"].diff() df["存栏量"]=df["存栏量"].dropna() def cfmapandsave(): for i in listlabel: df[i].plot() plt.legend() plt.show() df.to_csv('datacf.csv',encoding='gbk') def chafenacfpacf(): for i in listlabel: diff=df[i] diff=diff.diff() diff=diff.dropna() dftest = sm.tsa.adfuller(diff,autolag='BIC') dfoutput = pd.Series(dftest[0:4], index=['Test Statistic','p-value','Lags Used','Number of Observations Used']) print(i,"一阶差分 检验结果：") print(dfoutput) fig = plt.figure(figsize=(10,5)) ax1 = fig.add_subplot(211) fig = sm.graphics.tsa.plot_acf(diff, lags=20, ax=ax1) ax2 = fig.add_subplot(212) fig = sm.graphics.tsa.plot_pacf(diff, lags=20, ax=ax2) plt.subplots_adjust(hspace = 0.3) plt.show() diff.to_csv('datacf.csv',encoding='gbk') diff=diff.diff() diff=diff.dropna() dftest = sm.tsa.adfuller(diff,autolag='BIC') dfoutput = pd.Series(dftest[0:4], index=['Test Statistic','p-value','Lags Used','Number of Observations Used']) print("存栏量二阶差分 检验结果：") print(dfoutput) fig = plt.figure(figsize=(10,5)) ax1 = fig.add_subplot(211) fig = sm.graphics.tsa.plot_acf(diff, lags=20, ax=ax1) ax2 = fig.add_subplot(212) fig = sm.graphics.tsa.plot_pacf(diff, lags=20, ax=ax2) plt.subplots_adjust(hspace = 0.3) plt.show() def p_valuetest(): dftest = sm.tsa.adfuller(df['仔猪价格'],autolag='BIC') dfoutput = pd.Series(dftest[0:4], index=['Test Statistic','p-value','Lags Used','Number of Observations Used']) print("仔猪价格 检验结果：") print(dfoutput) dftest = sm.tsa.adfuller(df['玉米市场价格'],autolag='BIC') dfoutput = pd.Series(dftest[0:4], index=['Test Statistic','p-value','Lags Used','Number of Observations Used']) print("玉米市场价格 检验结果：") print(dfoutput) dftest = sm.tsa.adfuller(df['通货膨胀率CPI'],autolag='BIC') dfoutput = pd.Series(dftest[0:4], index=['Test Statistic','p-value','Lags Used','Number of Observations Used']) print("通货膨胀率CPI 检验结果：") print(dfoutput) dftest = sm.tsa.adfuller(df['鸡肉价格'],autolag='BIC') dfoutput = pd.Series(dftest[0:4], index=['Test Statistic','p-value','Lags Used','Number of Observations Used']) print("鸡肉价格 检验结果：") print(dfoutput) dftest = sm.tsa.adfuller(df['存栏量'],autolag='BIC') dfoutput = pd.Series(dftest[0:4], index=['Test Statistic','p-value','Lags Used','Number of Observations Used']) print("存栏量 检验结果：") print(dfoutput) org=pd.read_csv("data2.csv",encoding='gbk') org['时间'] = pd.to_datetime(org['时间']) org.index =org['时间'] org = org.resample('3M').mean() df=pd.read_csv("datacf.csv",encoding='gbk') df['时间'] = pd.to_datetime(df['时间']) df.index = df['时间'] df = df.resample('3M').mean() def pidchoose(): print('最优模型筛选') num=3 best_aic = float('inf') results = [] for i in range(num): for j in range(num): model = pf.ARIMAX(data=df,ar=i, ma=j,formula='猪肉价格~仔猪价格+玉米市场价格+鸡肉价格+存栏量',family=pf.Normal()).fit() aic = model.aic if aic < best_aic: best_model = model best_aic = model.aic results.append([i,j, model.aic]) results_table = pd.DataFrame(results) results_table.columns = ['ar','ma','aic'] best_model.summary() model = pf.ARIMAX(data=df,ar=0, ma=2,formula='猪肉价格~仔猪价格+玉米市场价格+鸡肉价格+存栏量',family=pf.Normal()) x=model.fit() print(x) x.summary() model.plot_fit(figsize=(15,10)) model.plot_predict(h=5, oos_data=df.iloc[-5:], past_values=73, figsize=(15,5)) # a=model.predict(h=5, oos_data=df.iloc[-5:])