A dynamic recommendation method based on user interest adaptive evolution

Abstract

The invention discloses a dynamic recommendation method based on user interest adaptive evolution, which comprises the following steps: firstly establishing a user interest model of each user, then segmenting user interest vectors according to a time sequence, calculating the similarity among the users, screening out a plurality of to-be-recommended users, and fully considering the characteristicthat the user interest is changed along with time; Based on historical comment records of the target user and the to-be-recommended user, obtaining a recommendation result; constructing a historical interest sequence matrix His; and performing interest evolution iteration for preset times through a PSO (Particle Swarm Optimization) algorithm to obtain a recommendation priority sequence of the target user and each user in the to-be-recommended user set, and finally recommending the users which are screened out from the to-be-recommended user set and ranked in front to the target user. By the adoption of the technical scheme, the stability of the recommendation result is controlled, meanwhile, the interest change of the user is reasonably simulated, and the recommendation accuracy, predictability and comprehensiveness are improved.

Description

Technical field [0001] The invention relates to the field of computer technology, in particular to a dynamic recommendation method based on the adaptive evolution of user interests. Background technique [0002] The recommendation system changes the passive mode of search engines to the mode of actively pushing information to users, while also providing users with a different, convenient and comfortable experience and effective information. Most current recommendation models tend to recommend resources that meet their interests, preferences or needs, such as web pages, movies, commodities, tags, and so on. Since the recommendation system of each platform is designed according to the characteristics of their respective platforms, they are very different in terms of their recommendation targets and purposes. Therefore, it is necessary to design suitable friend recommendations based on the actual situation and characteristics of the question-and-answer social network Algorithm. [0...

AI Technical Summary

Problems solved by technology

The recommendation system dominated by the interest similarity between users rarely considers the timing of user interests and ignores the interest changes of the recommendation subject, resulting in some blind spots or mistakes when recommending users, reducing the accuracy of recommendation

For example, for someone who currently has the same interest A as you, for some reason, his enthusiasm for interest A continues to decline over time, but the existing recommendation methods still give him a high recommendation priority, which reduces the accuracy of the recommendation

Images