Video recommendation method and device

Abstract

The embodiment of the invention provides a video recommendation method and device, and the method comprises the steps: extracting one or more video feature items from video data, and writing the video feature items into a video feature database; obtaining a configuration file of one recommendation mode, and determining one or more corresponding video feature items in the video feature database according to the configuration file; respectively calculating characteristic scores corresponding to the determined video feature items, and obtaining the total score of the video data; carrying out the ordering of all video data according to the total score, and generating a video recommendation list; and recommending the video recommendation list to a user. According to the embodiment of the invention, the videos recommended for the user meets the demands of the user. When the user wants to observe the favorite videos, the user can quickly obtain the favorite videos from the recommended videos, and does not need to carry out a large amount of search, thereby preventing server resources and network resources from being wasted severely.

Description

technical field [0001] The present invention relates to the field of multimedia technologies, in particular to a video recommendation method and a video recommendation device. Background technique [0002] In recent years, with the increasing popularization of network technology, large-scale video websites have appeared, and more and more people choose to watch videos online. [0003] Generally, when most users watch videos on a video website, they prefer to select newly released and popular videos. In traditional video websites, when users need to search for newly released and popular videos, they often need to enter keywords to search. When the server receives the user's search request, it needs to search the video library corresponding to the video and present it to the user. The user needs to find the desired video among the many presented videos. When the presented video does not have the video the user needs, the user needs to replace other keywords to search again un...

AI Technical Summary

Problems solved by technology

[0005] On the one hand, the number of views requires the accumulation of historical data, which can only be obtained after a period of time when the video is released. Moreover, the model predicts that when calculating massive video data, there will be a problem of cold start, which makes it impossible to release new releases in the first place. On the other hand, because the videos recommended to users are only analyzed from a single dimension of viewing quantity, the recommended videos often cannot meet the user's preferences. When users need to watch their preferred videos on video websites, It is still necessary to perform a large number of search operations by constantly changing keywords, and the server performs a large number of video search and presentation operations to obtain the desired video, and there is still a problem of massive consumption of server resources and network resources

Images