3 ms·
Personally, I think there is a lot of room for algorithmic improvement on recommendations. Most of the existing systems are really prediction algorithms, which
by nkurz 16y ago
Personally, I think there is a lot of room for algorithmic improvement on recommendations. Most of the existing systems are really prediction algorithms, which have been shoe-horned into making recommendations. I think this is primarily because accuracy of prediction is easier to measure (RMSE) than accuracy of recommendation, and the academic papers need something they can measure.
I agree that inferring 'like' is also difficult, and definitely prefer explicit rating systems. But for a site like Amazon, it's often a reasonable presumption that a user who buys an item likes it. But inferring that they dislike all the items they don't purchase is much less reasonable. Same for viewing webpages, or listening to music, or renting movies. If you want to be sure of a 'dislike', it probably will require action from the user.