How to choose Neural Net Architecture

Hello,

I'm reasonably new to Machine Learning and Neural Networks. I have a system at the moment which uses a Multi-layer Perceptron with Momentum Back Propagation as the learning rule.

A brief description of the task: Produce a ranked list of the top ten books to order in next year, based on things like number of loans, genre popularity etc.

My question is how do I know what the best architecture to use for this problem is? Also is there a name for this kind of problem that would give me better search results.

Thanks