Content recommendation method and device

Abstract

The invention relates to a content recommendation device which comprises a client and a server, wherein the client comprises a user login module, a user search information recording module and a display module; the server comprises a content push module and a user information data recording module. When individual contents are pushed, a simple recommendation method based on contents is improved, according to corresponding relationships between counted contents and browsing times, preference degrees of users about different contents can be calculated in real time, then contents can be pushed according to the preference degrees, and relatively precise pushing can be achieved; meanwhile, to develop potential content users, contents with similar habit preference are simultaneously pushed; a hot information display module is additionally set up ultimately, users can be assisted to pay attention to hot information, the view of users can be widened, and the users can be prevented from being addicted to the individual contents. The invention further provides a content recommendation method.

Description

technical field [0001] The present invention relates to a method and device for content recommendation. Background technique [0002] With the increase of Internet information, users generally need to retrieve the information they need. Due to differences in retrieval levels, it is often difficult for some users to obtain useful information quickly and effectively. At present, many user terminals push content based on user interests through content recommendation. For example, Chinese invention patent CN106202131A discloses a news recommendation method based on user interest; The recommendation method and system mainly weight the feature items of the user rating matrix according to the established query-multimedia classification matrix to obtain an improved user rating matrix, and then combine the user similarity based on the user query vector and the user rating matrix based The user similarity is calculated to obtain the overall similarity of users. However, the existing...

AI Technical Summary

Problems solved by technology

However, the existing content recommendation method still has the following defects: First, the existing algorithm is still based on the content recommendation method or collaborative filtering method. The analysis carried out, because many searches by users do not have statistical significance, and they don’t want to read it again after viewing it once. After counting these contents, it will affect the judgment of the content that the user really cares about, resulting in inaccurate push content; secondly, the existing algorithm is only based on user Searching content for recommendation, on the one hand, limits the opportunities to develop potential users; on the other hand, pushing the same type of content to users will cause users to see the information they want to see every day, and gradually deviate from the judgment of the outside world , which greatly magnifies the side that users want to care about, so that what users see is not a real world, leading to gradually extreme or even narrow thinking

Images