Video scoring method and device and video system based on film watching behavior of user

Abstract

The invention discloses a video scoring method and device and a video system based on film watching behavior of a user. The method comprises the steps of counting relevant indexes of all videos within preset time, and cleaning data of the relevant indexes and conducting mapping and converting on the data respectively; obtaining a score of each video according to preset weight, calculating an average score of videos within a scope of statistics and of one affiliation category according to the score of the video as a valid score of the videos, and calculating a valid score of videos which are not clicked out of the scope of statistics according to the affiliation category. According to the video scoring method and device and the video system based on film watching behavior of the user, firstly, various kinds of implied preference operation of the user to the videos are made full use of, the videos which are not clicked within certain time are given certain scores as well, and accordingly all the videos are scored; with the help of video relevance of albums and producers, preference characteristics of a client for actors, the producers, directors and the like are utilized, the accuracy of scoring is improved, and the precision of subsequent recommendation systems is improved.

Description

technical field [0001] The present invention relates to the technical field of video recommendation, in particular to a video scoring method, device and video system based on user viewing behavior. Background technique [0002] Although the current mainstream websites have designs for users to rate the videos they watch, most people in China have not developed the habit of rating and evaluating videos after watching them. Basically, the evaluation is obtained through the user's implicit operation on the video, which makes it impossible for the video website to accurately obtain effective information on the user's preference for the video. Moreover, the existing implicit operation evaluation mechanism is based on statistics for a period of time each time. Watching videos in the website, these watching videos may only account for a part of all videos, and the scoring system cannot cover all the videos, resulting in a greatly reduced effect of personalized recommendations desig...

AI Technical Summary

Problems solved by technology

[0002] Although the current mainstream websites have designs for users to rate the videos they watch, most people in China have not developed the habit of rating and evaluating videos after watching them.

Basically, the evaluation is obtained through the user's implicit operation on the video, which makes it impossible for the video website to accurately obtain effective information on the user's preference for the video. Moreover, the existing implicit operation evaluation mechanism is based on statistics for a period of time each time. Watched videos in the , these watched videos may only account for a part of all videos, and the rating system cannot cover all videos, resulting in a greatly reduced effect of personalized recommendations designed for users

Images