Study on collaborative filtering recommendation algorithm based on prediction for item rating

As the Internet is developing rapidly and people are becoming increasingly dependent on e-commerce, whether the users can be provided with accurate recommendation information has become the key to attracting users. Under this background, corresponding personalized recommendation systems have been proposed, and collaborative filtering algorithm is one of the algorithms commonly used in the recommendation systems. Collaborative filtering recommends the information in which the user is interested through analyzing the preferences of a group with similar interests and common experience. With the increasing number of users and items in the e-commerce system, the collaborative filtering algorithm is limited by data sparsity in operation, which will influence the quality of the recommendation results. This paper studied the collaborative filtering recommendation algorithm based on the prediction for item rating, and enhanced the accuracy of the recommendation algorithm by designing an improved similarity-based algorithm for predicting ratings based on the Pearson correlation coefficient.