I need a (simple) implementation of a regular feed-forward neural network in Python (from scratch, not a package).
I want an implementation from a Statistical point-of view, what I mean is an implementation like in the book of Tsay (2010), Section 4.19, where neural networks are described/implemented as an Econometric model. Or as described in Cheng and Titterington (1994). All neural nets have a non-linear regression representation, as far as I am aware.
So basically a simple neural network (you can look at it as a non-linear regression), where you have the pseudo-likelihood (Gaussian objective function), and then afterwards a numerical minimization with respect to the parameters of the network.
11 freelancers are bidding on average $38 for this job
I have already implemented it in C++ and python from scratch for MNIST data and trained it and got good accuracy Mesaage me in chat so that we can discuss more.