What is the architectural difference between using `feedforwardnet` and `fitcnet`?

I'm trying to train and evaluate a simple ANN for binary classification. It seems that I can use both fitcnet and feedforwardnet to do so and get classification probabilities. The architecture for fitcnet is described in the documentation here. It seems that the architecture is the same for feedforwardnet and that I can add additional layers for both. Is it true that their architectures are the same and that I can use both for binary classification, or is feedforwarnet used to solve other types of problems?

Answer 1

[Hiro Yoshino]

feedforwardnet is for regression problems whereas fitcnet is for classification problems.

[Thomas Kirsh]

Ah, thank you. I found this documentation link very helpful for finding more options for classification, as well as regression and other types of problems.

[Hiro Yoshino]

My pleasure.