Multi-media file recommendation method and recommendation server

Abstract

The invention provides a multi-media file recommendation method and a recommendation server. The method comprises the following steps that the recommendation server obtains N pieces of multi-media playing behavior information of N users, wherein N is a positive integer; the recommendation server clusters the N users according to the N multi-media playing behavior information so as to obtain M user groups, wherein M is a positive integer and is smaller than N; the recommendation server obtains multi-media playing behavior information of users to whom multi-media files are to be recommended, obtains user groups corresponding to the users to whom the multi-media files are to be recommended according to the multi-media playing behavior information, and recommends corresponding multi-media files to the users to whom the multi-media files are to be recommended according to the user groups. According to the multi-media file recommendation method, the problems that a multi-media file recommendation result is diverse, and the recommended multi-media files are not uniform in style and are not consistent with the preference of the users to whom the multi-media files are to be recommended are solved, the recommendation result is more stable, the influence caused by data sparsity and user playing behavior arbitrariness on the recommendation result is small, and user experience is improved.

Description

technical field [0001] The invention relates to the technical field of communication, in particular to a method for recommending multimedia files and a recommending server. Background technique [0002] At present, when recommending multimedia files (such as music) to users, a collaborative filtering method based on user similarity calculation is used to generate a recommendation list of multimedia files for the user to be recommended. users, and then generate a recommendation list for the user to be recommended through the playback preferences of K users. [0003] The current problem is that, on the one hand, the recommendation list of multimedia files generated by the collaborative filtering method based on user similarity calculation is unstable, which is greatly affected by data sparsity and the randomness of the user's playback behavior; The playback preferences among users are not necessarily similar, resulting in inconsistent style of the recommendation list generate...

AI Technical Summary

Problems solved by technology

This method solves the problems of divergent recommendation results, inconsistent styles, and inconsistency with user preferences when recommending multimedia files, making the recommendation results more stable, less affected by data sparsity and user playback behavior, and improving user experience.

Images