未找到预测包中的R forecast.holtwinters

时间:2017-07-28 13:28:48

标签: r forecasting holtwinters

我正在尝试使用forecast.holtwinters函数,当我尝试运行它时:

dftimeseriesforecast <- forecast.HoltWinters(data, h=65)

我收到此错误:

  

错误:找不到功能&#34; forecast.HoltWinters&#34;

我也试过这个:

 dftimeseriesforecast= forecast::forecast.HoltWinters(data, h=65)

但我收到此错误消息:

  

错误:&#39; forecast.HoltWinters&#39;不是来自&#39;命名空间的导出对象:预测&#39;

我使用以下代码查看预测包中的这个函数列表:

ls("package:forecast")

然后返回:

  

[1]&#34;%&gt;%&#34; &#34;准确性&#34; &#34; ACF&#34; &#34; ARFIMA&#34; &#34;马&#34; &#34; arima.errors&#34; &#34; arimaorder&#34; &#34; auto.arima&#34;
   [9]&#34;自动涂层&#34; &#34; baggedETS&#34; &#34;蝙蝠&#34; &#34; bizdays&#34; &#34; bld.mbb.bootstrap&#34; &#34; BoxCox&#34; &#34; BoxCox.lambda&#34; &#34; CCF&#34;
  [17]&#34; checkresiduals&#34; &#34;克罗斯顿&#34; &#34; CV&#34; &#34; CVAR&#34; &#34; dm.test&#34; &#34; dshw&#34; &#34;复活节&#34; &#34; ETS&#34;
  [25]&#34;发现频率&#34; &#34;预测&#34; &#34; forecast.ets&#34; &#34;傅立叶&#34; &#34; fourierf&#34; &#34;气体&#34; &#34; geom_forecast&#34; &#34; GeomForecast&#34;
  [33]&#34; getResponse&#34; &#34; ggAcf&#34; &#34; ggCcf&#34; &#34; gghistogram&#34; &#34; gglagchull&#34; &#34; gglagplot&#34; &#34; ggmonthplot&#34; &#34; ggPacf&#34;
  [41]&#34; ggseasonplot&#34; &#34; ggsubseriesplot&#34; &#34; ggtaperedacf&#34; &#34; ggtaperedpacf&#34; &#34; ggtsdisplay&#34; &#34;金&#34; &#34;霍尔特&#34; &#34; HW&#34;
  [49]&#34; InvBoxCox&#34; &#34; is.acf&#34; &#34; is.Arima&#34; &#34; is.baggedETS&#34; &#34; is.bats&#34; &#34; is.constant&#34; &#34; is.ets&#34; &#34; is.forecast&#34;
  [57]&#34; is.mforecast&#34; &#34; is.nnetar&#34; &#34; is.nnetarmodels&#34; &#34; is.splineforecast&#34; &#34; is.stlm&#34; &#34;毫安&#34; &#34; meanf&#34; &#34; monthdays&#34;
  [65]&#34; msts&#34; &#34; na.interp&#34; &#34;幼稚&#34; &#34; ndiffs&#34; &#34; nnetar&#34; &#34; nsdiffs&#34; &#34; PACF&#34; &#34;其余&#34;
  [73]&#34; rwf&#34; &#34; seasadj&#34; &#34;季节性&#34; &#34; seasonaldummy&#34; &#34; seasonaldummyf&#34; &#34; seasonplot&#34; &#34; SES&#34; &#34; sindexf&#34;
  [81]&#34; snaive&#34; &#34; splinef&#34; &#34; StatForecast&#34; &#34; STLF&#34; &#34; stlm&#34; &#34; taperedacf&#34; &#34; taperedpacf&#34; &#34;泰勒&#34;
  [89]&#34; tbats&#34; &#34; tbats.components&#34; &#34; thetaf&#34; &#34; trendcycle&#34; &#34; tsclean&#34; &#34; TSCV&#34; &#34; tsdisplay&#34; &#34; TSLM&#34;
  [97]&#34; tsoutliers&#34; &#34; wineind&#34; &#34; woolyrnq&#34;

有谁知道发生了什么事?我以前用过这个,没有问题。我使用预测版本8.1。

5 个答案:

答案 0 :(得分:8)

这些内容都不在forecast包中。它们位于stats

> m <- stats::HoltWinters(co2)
> class(m)
[1] "HoltWinters"
> p = predict(m)
> pp = stats:::predict.HoltWinters(m)
> p
          Jan
1998 365.1079
> pp
          Jan
1998 365.1079

predict.HoltWinters是来自stats的未导出函数,只应在来自HoltWinters()的对象上调用。

forecast.HoltWinters是来自forecast的未导出函数,这意味着您需要三个冒号才能访问它。但是你应该永远不必这样做,因为当你在forecast的输出上运行HoltWinters()时应该会自动找到它:

> m <- stats::HoltWinters(co2)
> forecast(m)
         Point Forecast    Lo 80    Hi 80    Lo 95    Hi 95
Jan 1998       365.1079 364.7139 365.5019 364.5053 365.7105
Feb 1998       365.9664 365.5228 366.4100 365.2879 366.6449
[etc]

同样:

> forecast:::forecast.HoltWinters(m)
         Point Forecast    Lo 80    Hi 80    Lo 95    Hi 95
Jan 1998       365.1079 364.7139 365.5019 364.5053 365.7105
Feb 1998       365.9664 365.5228 366.4100 365.2879 366.6449
[etc]

答案 1 :(得分:5)

像这样使用:

forecast:::forecast.HoltWinters().

它会起作用。

答案 2 :(得分:1)

使用R v3.4.4和预测v8.2时,这对我有用:

hw <- stats::HoltWinters(data) forecast_data <- forecast(hw, h=65)

答案 3 :(得分:0)

使用:

forecast_data <-forecast(mydata #data name,h=56)

在更新您的r版本之后,它将起作用

答案 4 :(得分:0)

您可以尝试下面的代码,它将起作用。您不需要HoltWinters.forecast。

    dftimeseries.hw <- HoltWinters(data)
    dftimeseries.forecast <-forecast(dftimeseries.hw,h=65)