Research on personalized hybrid recommendation system

With the popularity of the Internet and increasingly diversiform commodity, the recommendation system as a common approach came into our daily life, which supports an assistant decision-making when we purchase something on the internet. The traditional recommendation system is based on the user's collaborative filtering algorithm, and Amazon proposed a collaborative filtering algorithm to achieve good results. Through the study of two kinds of traditional algorithms, this paper proposed a personalized recommendation system model based on users and items. Then experiments were performed on the MovieLens-100K data set and the results of the recommendation were analyzed. Compared with the traditional collaborative filtering algorithm, the accuracy has been improved.