Micro-video personalized recommendation algorithm based on social network credibility

Abstract

The invention provides a micro-video personalized recommendation algorithm based on social network credibility and belongs to the field of computer algorithms. The micro-video personalized recommendation algorithm includes the steps of firstly, calculating user deviation degree according to the difference of global credibility and local trust; secondly, adding a confidence degree factor into a traditional similarity calculation method; thirdly, using the dependence of trust on time to allow a trust network to dynamically evolve; fourthly, creating a dual network time domain evolution model (DNTDEM) comprising a user similarity network and a user trust network; fifthly, acquiring a brand new user trust network according to the DNTDEM; sixthly, using an LDA model to supplement recommendationcontent; seventhly, allowing predicted users to be similar to the emotion neighbors of the users, and using a minimum error square value to optimize the similarity. The micro-video personalized recommendation algorithm has the advantages that high-quality new-form user-generated content (UGC) can be effectively identified and recommended to appropriate users; influence of the subjective prejudices of other users on the recommendation content can be relieved, and high-quality recommendation content can be objectively provided for target users.

Description

technical field [0001] The invention relates to a personalized recommendation method for micro-videos under a double-network time-domain evolution model of user deviation degrees, and belongs to the field of computer algorithms. Background technique [0002] Recommendations have been widely used, whether in academia or in other domains such as business. For example, product recommendation in Amazon, movie recommendation in MovieLens, and web page recommendation in Google. With the rapid development of the Internet, social networks such as Weibo, blog, QQ, and e-commerce have emerged and flourished, making information sharing and knowledge exchange more convenient for users. However, with the increase in the amount of information, users cannot filter out the information that is beneficial to their needs in a short time in the face of huge data. At present, the research on recommendation algorithms of social networks has involved many platforms, including online social websi...

AI Technical Summary

Problems solved by technology

However, little research has been done so far on the recommendation of bit-sized videos

Secondly, in social media, there are also challenges in the personalized recommendation of high-quality UGC. Although some candidate items are easily recognized (such as through multiple "favorites"), for potential candidate items with little evidence , which is a long-tail problem

Therefore, a common question is whether it is possible to identify these high-quality UGCs and recommend them to appropriate users

Images