How to implement neural networks?

Question

[nabeelmamoon]

So I have an array of tissue samples, each with methylation values for a set of specific genes. These samples are classified as different types of cancers, with about 8 types in all. Is there any way I can use neural networks to predict the conditions of test samples?

Array like this:

How would I do this in R programming?

Answer 1

[NetunoPoncã]

You are talking about classification, that is, inducing a model based on training data in order to classify (predict the labels) of previously unseen data. Neural networks are just one possibility. If you want to implement them from scratch, I suggest you take a look here:

• Neural Networks and Learning Machines by Simon Haykin

Matlab has a neural network toolbox. You can also find examples in R, for instance: http://gekkoquant.com/2012/05/26/neural-networks-with-r-simple-example/