User personalized preference prediction method based on multi-angle non-transfer preference relationship

Abstract

The invention discloses a user personalized preference prediction method and device based on a multi-angle non-transfer preference relationship, and the method comprises the following steps: carryingout the modeling of user preferences through a multi-angle preference difference feature model, and building a user multi-angle non-transfer preference model meeting the personalized demands of a user; constructing a user scoring model corresponding to the user multi-angle non-transmission preference model, wherein the user scoring model is used for obtaining a score actually given by a user to atarget object; and achieving a preference score prediction model under multiple angles by integrating the user multi-angle non-transmission preference model and the user scoring model so as to calculate user preference score data. According to the method, the accuracy of user preference prediction can be improved in a personalized recommendation scene, and the effectiveness of a recommendation result can be verified.

Description

Technical field [0001] The present invention relates to a personalized recommendation method, in particular to a user personalized preference prediction method based on a multi-angle non-transitive preference relationship. Background technique [0002] In the Internet age, personalized recommendations have begun to be integrated into all aspects of human material and cultural life. Various graphics, text, audio and video content have gradually come to the era of targeted delivery with personalized recommendations as the goal. Such as "NetEase Cloud Music" online music playlist recommendation, "Meituan Dianping" restaurant food recommendation, "Xiaohongshu" beauty experience recommendation, "Tik Tok" fun short video recommendation, etc., these are hidden behind the app The intelligent system that understands people's minds has greatly enhanced users' enthusiasm for use, optimized users' consumption experience, and has been widely recognized by the entire society. [0003] Generall...

AI Technical Summary

Problems solved by technology

However, judging preferences between two items, the above also explained that the factors contained in it may be extremely complicated.

When evaluating from a single perspective, the predicted user preference results are likely to meet the transitive characteristics, but when evaluating from multiple perspectives at the same time, user preferences may no longer satisfy the transitive characteristics

Images