python predict_请教:python 时间序列模型中forecast()和predict()的区别

该楼层疑似违规已被系统折叠 隐藏此楼查看此楼

model=ARIMA(xdata,(p,1,q)).fit()

print(u'predition:',model.forecast(5))

输出结果:

(array([ 87375195.39200352, 87294371.90756984, 87334812.48326722,

87424680.24570116, 87534694.51144411]),

array([ 1062798.88935205, 1148529.23429549, 1162499.18483996,

1164944.86673523, 1165411.07431117]),

array([[ 85292147.84606433, 89458242.93794271],

[ 85043295.97315933, 89545447.84198035],

[ 85056355.94892374, 89613269.0176107 ],

[ 85141430.2629253 , 89707930.22847703],

[ 85250530.77861008, 89818858.24427813]]))

predict_fig1 = model.predict('2014-11-11','2014-11-16',dynamic=True)

print(predict_fig1)

2014-11-11 -658782.778083

2014-11-12 -195135.421861

2014-11-13 -6152.928357

2014-11-14 70876.267154

2014-11-15 102273.339185

2014-11-16 115070.774308

Freq: D, dtype: float64

这两个方法都是做预测,但输出结果不同,到底有什么区别?

官网解释如下,仍不明白,求解释~

http://statsmodels.sourceforge.net/stable/generated/statsmodels.tsa.arima_model.ARIMAResults.predict.html#statsmodels.tsa.arima_model.ARIMAResults.predict

你可能感兴趣的:(python,predict)