I see you keep editing, so I'll respond to the latest version of your comment: Assuming an ML-based methodology (and I would not personally assume that), yes,

I see you keep editing, so I'll respond to the latest version of your comment: Assuming an ML-based methodology (and I would not personally assume that), yes, the training data, the resultant weights, and the algorithm itself would be of interest, though at least the last two would be necessary.