Computing feature vector for user and item/movies for content-based filtering

Hello DeepLearning.AI community.

I cannot get my head around with why are computing feature vector for users and items. And also why are calculating it with NN and not with other model?

Kindly explain in a beginner-friendly way.

Thank you.

Hello @annoyingCode,

Because with them, we can compare between an user and an item by computing the similarity of their feature vectors. If an item is “similar” to an user, then we can recommend that item to the user.

Because NN can learn feature vectors by maxmizing the “similarity” of an user and an item that is highly rated by the user. Decision tree can’t.