Multivariable forecasting in AI Studio

iceman0723
iceman0723 Altair Community Member
in Community Q&A

I am working in a multivariable sales forecasting in AI Studio. I approached it in 2 different manners: first taking all data (which contains a KEYS columna that relate data from a 5 year period of time, the "KEY" column represents a distinct time series. This means that the data associated with each key is independent in terms of its temporal progression) and then trying to manage KEYS column as a multivariate. In the second case, I could not find a result. How to perform multivariable forecasting (with different keys and periods of time that are not unique) in the right way?

Tagged:

Answers

  • Garima_Singh
    Garima_Singh
    Altair Employee
    edited February 20

    Hi @RolandJones, FYI. Kindly provide the required support.

  • RolandJones
    RolandJones
    Altair Employee

    Hi @iceman0723

    Are you able to share part of the data? I should be able to build a small example. As I understand it, you have multiple separate time series datasets which have been appended, and can be identified by a KEY column. Please correct me if I'm wrong.

    My suggestion would be to process each time series independently, and try lagging the predictive variables. I would then append the data back together and try applying a predictive model.

    Best,

    Roland

Categories