主要讲四部分内容：
non-personized systems
popularity: 基于流行度或者最大利益化的推荐。 缺点也明显：你可能在特殊地方有些特殊需求， 或者你本来就是大多数人不一样
Association:
collaborative filtering
matrix factorization (and its variant like probablistic matrix factorization), also known as SVD
Deep learning
[Udemy] Recommender Systems and Deep Learning in Python
原文：https://www.cnblogs.com/mashuai-191/p/10861192.html