A method to reduce Matthew effect in movie recommendation system

Abstract

The invention discloses a method for reducing the Matthew effect in a movie recommendation system, which comprises the following steps: obtaining a movie scoring matrix and a movie popularity set; establishing a Matthew effect-affected movie set; and calculating the Matthew effect-affected movies in the movie set Popular periodic function and standard popular periodic function; based on the popular periodic function and standard popular periodic function of each movie in the movie collection affected by the Matthew effect, calculate the residual square sum of each movie in the movie collection affected by the Matthew effect, and establish a residual set ; Based on the set of residuals, the average score of each movie in the Matthew effect affected movie collection is punished, and the new average score of each movie in the Matthew effect affected movie collection is obtained, thereby reducing the Matthew effect in the movie recommendation system. By adopting the method disclosed by the invention, the Matthew effect in the movie recommendation system can be reduced, and the situation that popular movies have high popularity for a long time and unpopular items are not paid attention to by no one can be prevented.

Description

technical field [0001] The invention relates to the field of movie recommendation systems, in particular to a method for reducing the Matthew effect in the movie recommendation system. Background technique [0002] The recommendation system is a tool to help users quickly discover useful information. It is mainly characterized by modeling the user's interest by analyzing the user's historical behavior, so as to actively recommend information to users that can meet their interests and needs. The movie recommendation system is one of the important components in the field of recommendation systems. It collects information about users and movies to help users quickly find movies that interest them in a huge movie library. A good recommendation system not only requires High recommendation accuracy and recommendation coverage should also have better performance in anti-cold start ability and ability to overcome matrix sparsity. [0003] The Matthew effect, that is, the so-called ...

AI Technical Summary

Problems solved by technology

[0003] The Matthew effect, that is, the so-called effect that the strong are stronger and the weak are weaker. If it is a recommendation system, it will increase the popularity gap between popular items and non-popular items, making the popularity develop in a polarized trend, making it difficult for users to find out what they really are. Items of interest, reducing the coverage and accuracy of the recommendation system, which will have a very negative impact on the recommendation system

The current mainstream recommendation algorithms all have the Matthew effect

Images