Video recommendation system

Abstract

The invention relates to a video recommendation system. At least two HTTP servers http-server serve as a system access entry, the recommendation server cluster consists of recommendation servers; internal logic of each recommendation server is divided into a data dispatcher, a data recall device, a data filter, and two historical servers adopting a main / standby mode and are divided into two typesof data, namely data in the model, real-time data and two recall pool servers, the main / standby mode is adopted, a data caching function is achieved, the data recall device stores the recall data intothe recall pool servers, and the data filter obtains the data from the recall pool servers. According to the invention, the system architecture is clear, the processing efficiency is high, the stability is good, the problem of insufficient system resources can be effectively solved, video recommendation and audio recommendation functions can be realized, the user experience is improved, repeatedpushing can be avoided, and the affinity of the pushing experience is improved.

Description

Technical field [0001] The present invention relates to the technical field of multimedia content recommendation based on user behaviors, in particular to a video recommendation system. Background technique [0002] In the era of rapid development of multimedia technology, the integration of music and short videos has become a new trend of development. Based on the "music + social" approach, more and more music apps have begun to try and provide short video functions. +Short video" to further enrich the entertainment of the music APP, further enrich the content quality and richness of the APP friends dynamic circle, and enhance the user activity of the APP friends dynamic circle. [0003] The music + short video approach involves B2C (Business-to-Customer) business, which is characterized by large traffic and large amounts of data. At the same time, a large number of users will have high concurrent access to the system, which causes a lot of The access pressure is extremely prone ...

AI Technical Summary

Problems solved by technology

[0003] The method of music + short video involves B2C (Business-to-Customer) business, which is characterized by large traffic and large amount of data. At the same time, a large number of users will have high concurrent access to the system, which will cause a lot of damage to the system. The access pressure is extremely prone to problems of insufficient system resources, such as insufficient memory, cpu, and network bandwidth on the server side, etc.

Images