Signed Network Recommendation Method with Incorporated Negative Link Information
XUE Miaomiao
WANG Wei
Abstract：To alleviate the problem of low recommendation performance due to sparse data,a signed network recommendation method fused with negative link information is proposed.This method introduces the negative link information,comprehensively con-siders the sign information in the network.Firstly the positive and negative feature information of the link is extracted from the net-work,that is,the features related to the relationship between users,and finally the sign features of the link with the recommendation is combined with the behavioral features in the model,the relationship between trust and distrust is obtained by using logistic regres-sion,and it is integrated into several representative social recommendation algorithms to explore the performance of score prediction.This paper conducts experiments based on the extended Epinions dataset.Compared with other recommendation algorithms,the method proposed in this paper has better performance in MAE and RMSE indicators,and achieves the purpose of improving the rec-ommendation performance.
Keywords：signed networkrecommendationfeature
Publication Date：2023-11-20
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：7( 2580-2585,2596 )