Method for improving diversity of personalized recommender systems

Abstract

The invention discloses a method for improving diversity of personalized recommender systems. The method specifically comprises the following steps of: firstly obtaining a score data set of a user; carrying out recommendation on the score dataset by using three frequently used recommendation algorithms; carrying out threshold value control and sorting on recommendation results; and displaying a final result to the user. Through re-sorting the recommendation results, the list content of the final recommendation result can be changed, so that the diversity of the systems can be improved; meanwhile, a threshold value control method is used during the sorting, so that the recommended articles are liked by the user as far as possible and the prediction correctness of the recommender systems is ensured. Experimental results show that compared with the existing personalized recommender system method, the method system disclosed by the invention can greatly improve the diversity of the systems under the condition of less reducing the prediction correctness, thereby achieving the aim of controlling the system correctness and diversity balance.

Description

technical field [0001] The invention belongs to the field of Internet, mobile Internet and computer network, and more specifically relates to a method for improving the diversity of a personalized recommendation system. Background technique [0002] Due to the long-tail effect in the field of online retailers, that is, 80% of the sales of merchants come from 20% of their products, and if the sales of long-tail items of retailers can be improved, then the business of online retailers The amount will increase exponentially and can meet the unique preferences of users. Therefore, in recent years, the personalized recommendation system combined with computer network technology and big data processing technology has gradually attracted people's attention in the field of e-commerce and has been widely used. The personalized recommendation system is to realize the information filtering based on the user's personal interests. The personalized recommendation system is an information...

AI Technical Summary

Problems solved by technology

[0006] In view of the above defects or improvement needs of the prior art, the present invention provides a method for improving the diversity of the personalized recommendation system, the purpose of which is to solve the existing methods that do not consider the diversity of the personalized recommendation system , resulting in technical problems of unsalable long-tail items of Internet retailers

Images