The company's new "Netflix Prize" will go to "the first person who can achieve certain accuracy goals in recommending movies based on personal preferences." While the idea has its detractors, it's getting quite a bit of attention. Read more here.

Somehow this doesn't seem feasible.

Much of how well you like any given movie is determined by external influences, especially mood at the time the movie began.

Personally, their recommendation system works just fine for me, though every once in a while it'll suggest I watch Speed 2.

I agree Lerza this is going to be hard, but all they are looking here is some good publicity and the possibility of a better algo or improvement to the system. So I say why not do it. What data is being offered in the download?

member
movies rented
movies ratings
ratings by the member

?

Does the data come with only a member ID? It would seem if you were able to able to bring the members personal info into the mix it would work much better. Another possibility is factoring in the directors, producers, actors/actresses personal preference and relationships within the industry. If I like Quentin Tarantino then I will probably like some of his favorite movies to. Just going outside the membership info and utilizing public relationships would be the first major step here.

__________________
KBK Marketing | Jaan Kanellis Search Marketing Blog | Be the best and a link is what you'll get!