I have been tasked with combining a several classifier models we have into one model using deep learning (or something else).
The reason for this is that, in future, it would be difficult to maintain 20-30 models separately but rather have one 'master' model that can predict for all targets. This new 'master' model will have say 4 targets, so would expect 4 probabilities as outputs. All current models have the same features.
This is illustrated using the toy example:
The model has 6 features and 4 targets (is dog, is cat, likes making noise, is good boy/girl)
The idea is that there is now one model with 4 outputs where previously there were 4 separate models.
What would be the best method to achieve this? It doesn't have to been neural networks but it was mentioned as a solution.