training neural network in R using costfunction and gradientdescent

Vote count: 0

I have implemented a neural network assignment in Matlab and I'm trying to make it work in R too. I have already created the costfunction and gradientdescent which have the same results both in Matlab as in R so I expect them to be correct. In Matlab quite a complicated optimization script is used to train the network. I was wondering if there is a way to train the model with the optim-function (or something equivalent) in R? I have tried it already with a couple of methods but the outcome is different resulting in a lower accuracy. Does anyone have any suggestions how to do this without using a package dedicated to neural networks?

Thanks, Paul

asked 1 min ago

Polhovsky

This entry passed through the Full-Text RSS service - if this is your content and you're reading it on someone else's site, please read the FAQ at http://ift.tt/jcXqJW.

training neural network in R using costfunction and gradientdescent