using own dataset

[KKindom]

With some pretty simple changes we can do things like read N + k variables (with k additional sources of information) and still only predict N.

Could you elaborate more what these simple changes are? Would they work in a setup where N is unavailable?