search

robjhyndman / forecast

Forecasting Functions for Time Series and Linear Models
http://pkg.robjhyndman.com/forecast
1.11k stars 341 forks source link

Stepwise for selection of covariates #856

Closed betmar78 closed 4 years ago

betmar78 commented 4 years ago

The auto.arima function applies stepwise presenting the best regression model with ARIMA errors when the xreg parameter is used. However, it does not select the most significant covariates; presenting all covariates, whether significant or no.

Is it also possible to apply stepwise to select the significant covariates (xreg) in the result presented by the auto.arima function for the best model?

How to define the final model with only the significant regressive variables (xreg)?

robjhyndman commented 4 years ago

It is possible, but not implemented. You can manually select the model with the smallest AICc value.

betmar78 commented 4 years ago

But this function does for ARIMA errors. It does not do for covariates. Thanks for the reply.