Comment by streetcat1

Can you elaborate more on why you choose those? . More specifically: Did you benchmark them using known datasets? (Against other algorithms?)

Also, where do you store the training data?