Dear ALL;
I have made a sales and demand forecasting autoML model. It trained well and is working. I am looking for a way to format input data for batch prediction forecasting where I would like to do more than one forecast horizon predictions.
Basically my model uses monthly data granularity with 12 month context length and 6 month forecast horizon. (I have 15 covariate features and predict for 14 separate identifiers).
What I would like to do is configure the input data for a batch forecasting where the forecasting would start 3 month earlier than where the feature data end and make forecasts for these three time periods AND when the feature data ends do the normal 6 month forecasting so I end up with 3 month forecasting where I know the actuals and the normal 6 month forecasting.
At this point no matter how I format my input data I only get the 6 month forecast horizon.
Thanks
I searched for a document which explain how batch prediction results work within Google Cloud, and I hope it helps you.
User | Count |
---|---|
2 | |
1 | |
1 | |
1 | |
1 |