i want to use MAPA in group time series modelling . I have temerature data of 20 weather stations , I want to forecast individual stations as well as average over the region . Please update me how to combine grouped time series of hts package and MAPA.
I think the easiest way to do it is to first forecast time series with MAPA separately, and then use the combinef function in hts package.
Also, you can get in-sample MSE from the MAPA output for the reconciliation, if you need it.
I want to use MAPA with different set of regressors on each level, including macroeconomic input on the monthly and annual level of aggregation of daily sales. Looked at the CRAN document, but did not find a place to feed in xreg parameters. Am i missing something?
It is not available for R yet. However an ETS implementation with regression inputs is now almost complete in the TStools package, so when I get the time I plan to build a MAPA with that in its core, which should be able to handle regressors as described here.
i want to use MAPA in group time series modelling . I have temerature data of 20 weather stations , I want to forecast individual stations as well as average over the region . Please update me how to combine grouped time series of hts package and MAPA.
I think the easiest way to do it is to first forecast time series with MAPA separately, and then use the combinef function in hts package.
Also, you can get in-sample MSE from the MAPA output for the reconciliation, if you need it.
Nicos,
I want to use MAPA with different set of regressors on each level, including macroeconomic input on the monthly and annual level of aggregation of daily sales. Looked at the CRAN document, but did not find a place to feed in xreg parameters. Am i missing something?
Dmytro
Hi Dmytro,
It is not available for R yet. However an ETS implementation with regression inputs is now almost complete in the TStools package, so when I get the time I plan to build a MAPA with that in its core, which should be able to handle regressors as described here.
Nikos
Really cool, thanks! Installed TStools and will investigate closer.
Have you an example in MAPAx with exogeneous variables (trivail example) . I try to do a benchmark of model but MAPAX doesn t work with my code.
Have a look at this post: http://kourentzes.com/forecasting/2017/11/09/mapax-example-for-r/
Hope this helps!