Forecasting using SAS Forecast Server, SAS/ETS, and more

Impute time series with missing info for y

Reply
Frequent Contributor
Posts: 126

Impute time series with missing info for y

Hi all,

I am facing the following problem: I have complete series for Y and X1 X2 X3 only for a period of 11 months (not enough to fit a model and to give good forecasts).

I could obtain an additional 12 month period back in time, only thing is i dont have any info about Y but the info for  X1 X2 X3 is there.

What would be the best way to impute values for Y?

Thanks in advance

SAS Employee
Posts: 416

Re: Impute time series with missing info for y

Need to edit my suggestion - as I only realized now, that you don't have any previous values for Y at all (note to self: reading skills are a plus).

In this case my suggestions below won't work of course.

Hello -

Time series models typically deal with missing values well, if there are only few. So you might want to try your data as is first.

As another idea, you may want to use an exponential smoothing model for your Y series first, and use its predictions to impute missing values of Y.

Thanks,

Udo

Ask a Question
Discussion stats
  • 1 reply
  • 572 views
  • 0 likes
  • 2 in conversation