1

I'm trying to use the Holt-Winters model included with the R package 'forecast' to forecast a product's sales revenues, which includes seasonality.

In the past the product's price as changed, and with it the revenues generated - events that I believe are referred to as 'explainable variation'.

My question is - how do I take into account these changepoints when forecasting? Is the general process to 'standardise' the historical data - by, say, applying the % increase/decrease in revenues brought about by a price change to revenues prior to the change - or introducing some kind of factor account for the % shift? Maybe there is an automated way in R to account for these events?

AndyL
  • 11
  • 2
  • you might want to take a look at this [post](http://stats.stackexchange.com/q/104882) – ab90hi Jan 12 '17 at 13:02
  • thank you, this is interesting! The aforementioned price changes are though not outliners - they do not need detecting as I know when the changes were made, and what the impact was – AndyL Jan 12 '17 at 14:19
  • Can you clarify whether or not you'll know when forecasting when these price change 'events' will occur in the future? – AnscombesGimlet Jan 19 '17 at 16:41
  • No, I will just know when specific changes were made in the past – AndyL Jan 20 '17 at 12:56

0 Answers0