I really want to try artificial neural network for stock trading except the programs I've found are too expensive and not customizable. So I'm planning to make one myself. But before I get too serious, how does artificial neural network perform compare to traditional algos for market prediction?

I would adopt a genetic basis to it too. Create survivability on capital appreciation. The inputs could be hundreds of data fields. Ie time, date, price, high low close, .. Day of week month.. Any data field that could effect price.. The visual component is each 'lifeform' would be the chart construct.. The analytical components would be statistical measures, or indicators.. Actions would be buy sell short cover.. The lifeforms that survive would be able to duplicate and survive for the next generation.

I was planning to use encog which is from heatonresearch. This is a very interesting article because it shows an actual implementation using the framework. But it seems like the wiki pages have been deleted for some reason..

Using the genetic algorithm for this project seems like a crazy and great idea. I didn't even think of it. As far as the input goes I'm not sure if it is a good idea to use every possible input because some of them may have near zero correlation and only increases the computation time. I guess it needs some pruning.

As Bowgett mentioned and (with my little experience) I also agree... I think neural nets are mostly useful for on-going parameter optimization vs using them to make trading decisions. What I do and suggest to others.... I would focus on building profitable strategies first then see where a NN could help with optimization if its needed. With as much computing power is available today, from Amazon EC2 and others, you should be able to complete optimization for hundreds of instruments by the end of the weekend for the coming trading week.

Sorry for not being clear. By traditional algos, I meant non-linear mathematical functions. Neural network is, as you know, a mathematical model for function approximation.