Dear All,
Let's say that I want to produce a model that will predict sells of month M based on the sells figures of the 2 previous months.
Let's say that I have trained this model on the last 3 years (in order to get the seasonning effect and so on...).
Let's say that I go live with my trained model on jan 2020.
In feb 2020, I have new data (actual jan 2020 sells) that I would like to feed to my model in order to enhance it (and to keep it up to date).
Is there any way to "keep training" my model with the feb 2020 new data? or do I need to fully re-train the model, based on all data?
Thank you for your reply!
Solved! Go to Solution.
@alts2h take a look at this article: https://community.alteryx.com/t5/Alteryx-Promote/Alteryx-Promote-vs-Concept-Drift-Data-Drift-Drift-M...
You could set up a workflow to monitor the drift or decay of a model and then trigger a retraining if it fell outside of acceptable results via the Alteryx Server.
Interesting suggestion...
I'll study this solution
Thank you!