Suppose there are data from three stores with the same attributes but different numbers of events. It is necessary to create a common prediction system, for example, sales for each store separately and in the general case. And find the validation value of the found common prediction system for each store separately. Whether there is help on the task?
P.S. I have one idea about this. Create an additional attribute - “store number”. In this case, use the filters and create a validation filtering by store number. How correct is this in your opinion?
I am not sure if I understand correctly what you are planning to do but let me try to help.
You want to create prediction models separately for each store and then an overall model for all stores.
Do you know how many stores you have? I guess you will need to learn for every store a new machine learning model. For this you can of course create a new variable to filter the data. I would not recommend to use the store number as a integer as this could have impact on your prediction model or at least leave the store number out of the machine learning model.
By the way…
If you want to check example workflows which include machine learning models or simply look for nodes for certain tasks I would warmly recommend the KNIME Community Workflow Hub: https://hub.knime.com/
OK thanks. Thanks for the advice