How to do an incremental learning where new items appears gradually?

[lmatejka]

I suppose that NVT workflow process whole dataset and set cardinatlity and labels of item_ids and category_ids? When I want to do a continuous learning day by day, I can train model on new interactions from that day, but when new item_id arrives, no label is present in NVT workflow. What is a proper solution for that? I have to perform new NVT transformation on whole dataset with all items included?