Ok guys this is a great programming challenge especially for you AI types. NetFlix is running a contest to come up with better heuristics for determining if a given person will like a given movie based upon certain data such as what movies they have liked in the past. The overall prize is $1,000,000 for a 10% increase over their current methodology but they are also giving away $50,000 incremental prizes. http://www.netflixprize.com/rules
The rather obvious answer would be forward feed ANN but I think one could do better by applying an associative memory in the feature detection process.
Anyone interested in creating a team effort?