Personalized recommendation method and system based on label information

Abstract

The present invention relates to a method and system for personalized recommendation based on label information. The method comprises: generating a three-dimensional table according to the state of marking the identification of each product according to the information of each label corresponding to the identification of each user, establishing a user product interaction matrix model and a product label relation matrix model according to the three-dimensional table; constructing a joint decomposition model corresponding to the user identification, the production identification and the label information according to the user product interaction matrix model and the product label relation matrix model; employing a Bayes personalized ranking method to solve the joint decomposition model to obtain a plurality of parameter values; obtaining the preference degrees of the user identification to the identification of each product according to the parameter values; and selecting the recommendation identifications from the identifications of the products according to the preference degrees, and recommending the information of the products corresponding to the recommendation identifications to terminals whether the corresponding user identifications are located. Therefore, the method and system for personalized recommendation based on the label information can solve the limitation of the data sparsity in the label information to improve the precision of the personalized ranking so as to improve the accuracy of the recommendation.

Description

technical field [0001] The invention relates to the field of information technology, in particular to a method and system for personalized recommendation based on tag information. Background technique [0002] With the development of information technology, e-commerce has developed rapidly, and the information overload phenomenon of e-commerce platforms has become more and more serious. To alleviate information overload, e-commerce platforms usually use personalized recommendation systems. The personalized recommendation system builds a user interest preference model based on the online browsing data or purchase data of individual users, so as to recommend products that meet their unique needs to users, which can optimize user experience and increase platform user traffic. [0003] The traditional personalized recommendation system recommends products with similar tags to users according to the tags marked by users on purchased or browsed products. However, users generally...

AI Technical Summary

Problems solved by technology

However, users generally mark products with few tags, and a product contains a limited number of tags. Therefore, the tag information included in the e-commerce platform is sparse, and the tag information cannot accurately represent the description of the product. At the same time, because the tag information The sparsity, the similarity between users is also inaccurate, resulting in low recommendation accuracy

Images