2

I am using auto.arima from forecast package for time series. The auto.arima selects best p,d,q parameters by itself so need not to worry. But I get worried when I see all forecasting values which are exactly same. I think it is not correct.

Also in a tutorial given below, they are getting same forecast values 67.75

Ref: http://a-little-book-of-r-for-time-series.readthedocs.org/en/latest/src/timeseries.html

> library("forecast") # load the "forecast" R library
> kingstimeseriesforecasts <- forecast.Arima(kingstimeseriesarima, h=5)
> kingstimeseriesforecasts
     Point Forecast    Lo 80    Hi 80    Lo 95     Hi 95
  43       67.75063 48.29647 87.20479 37.99806  97.50319
  44       67.75063 47.55748 87.94377 36.86788  98.63338
  45       67.75063 46.84460 88.65665 35.77762  99.72363
  46       67.75063 46.15524 89.34601 34.72333 100.77792
  47       67.75063 45.48722 90.01404 33.70168 101.79958

Please let me know what additional preprocessing should I perform before feeding to auto.arima so that I get correct forecast values?

Nick Cox
  • 48,377
  • 8
  • 110
  • 156
bioinformatician
  • 293
  • 2
  • 4
  • 9
  • 1
    Being puzzled by results is not evidence that they are wrong. Note that the intervals widen into the future as they should. – Nick Cox Jun 09 '15 at 07:57
  • 1
    According to the information in the link that you give, the ARIMA(0,1,1) model is fitted. The forecast function of this model is constant, see for example [this post](http://stats.stackexchange.com/questions/125909/forecast-using-arima-models). – javlacalle Jun 09 '15 at 08:20
  • Don't confuse a forecast (average values) with a simulation (average values + random noise). See also [ARIMA forecasting life expectancy](http://stats.stackexchange.com/q/68379/17230), [Difficulty with auto.arima function results](http://stats.stackexchange.com/q/84255/17230), [How do I forecast using exponential smoothing past one data point?](http://stats.stackexchange.com/q/124000/17230), & [Forecasting from different time horizon](http://stats.stackexchange.com/q/135068/17230). – Scortchi - Reinstate Monica Jun 09 '15 at 08:24

0 Answers0