Read all about it, via Brad DeLong. I’ll add a dose of Michael Polanyi to the debate and note that recommender systems work best when combined with non-articulable knowledge.
The Netflix prize, 300 days later
Previous post: Do pesticides contribute to autism?
Next post: Can a mosque be sexy? Yes.
It is pretty tough to create a good recommender system from a set of static data. But imagine you could change your recommendation system every hour, try it out on thousands of customers, and see what worked and what didn’t.
Trying to solve the recommendation problem based on a dataset is thinking like an economist. A learning system with trial and error feedback is thinking like a computer programmer.
I come from asia, injoy 室內è¨è¨ˆ,work in a æ¬å®¶å…¬å¸,other enjoy isecosway和科士å¨,good see U。
Thanks.
Comments on this entry are closed.