Video screening method and device, electronic equipment and storage medium

Abstract

The invention discloses a video screening method and device, electronic equipment and a storage medium. The video screening method comprises the steps that features of a user, features of a first video operated by the user and operation features of the first video are acquired; Processing the characteristics of the user, the characteristics of the first video and the operation characteristics of the first video by adopting a pre-trained screening model to obtain a target score corresponding to each first video; And screening out a preset number of first videos from the first videos operated bythe user according to the target score. The first video operated by the user is screened through the screening model, the trigger source quality of video recommendation can be ensured, and low-quality or invalid videos are filtered out, so that the video recommendation efficiency and fineness can be improved; And the trigger source can comprise the first video in a longer time range, so that thelong-term interest of the user can be described more effectively, the trigger recall time range is expanded, and the personalized recommendation effect is improved.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a video screening method, device, electronic equipment and storage medium. Background technique [0002] Video recommendation technology generally includes two steps of triggering and sorting. First, according to different trigger sources, various trigger strategies are used to obtain trigger results, and then the trigger results are sorted and screened through the sorting model, and the final recommendation results are displayed to users. Therefore, an accurate trigger source can more effectively meet the individual needs of users. [0003] Collaborative Filtering is commonly used as a trigger in traditional video recommendation systems. For example, item-based collaborative filtering (item-cf) uses the similarity between items for recommendation and ranking. In this item-cf-based triggering method, all behaviors such as the user's recent clicks, likes, and concerns...

AI Technical Summary

Problems solved by technology

[0004] However, in short video applications with a large number of videos and rich user behaviors, due to the limitation of the online system, the user behavior as the trigger source can only save the most recent limited part, resulting in the lack of ability of the trigger source to express long-term interests of users; Moreover, the user behavior as the trigger source has not been screened, and the overall trigger quality is difficult to guarantee

Images