基于改进的LFM算法的短视频推荐系统的研究与实现  

Research and implementation of short video recommendation website based on the improved LFM algorithm

作　　者：彭宇 宁慧[1] 张汝波 PENG Yu;NING Hui;ZHANG Rubo(College of Computer Science and Technology,Harbin Engineering University,Harbin 150001,China;College of Mechanical and Electronic Engineering,Dalian Minzu University,Dalian 116600,China)

机构地区：[1]哈尔滨工程大学计算机科学与技术学院,黑龙江哈尔滨150001 [2]大连民族大学机电工程学院,辽宁大连116600

出　　处：《应用科技》2022年第3期64-68,共5页Applied Science and Technology

基　　金：国家自然科学基金项目(61673084)。

摘　　要：为解决传统潜在因子模型(latent factor model,LFM)算法无法为用户实时推荐的问题,本文在LFM算法的基础上结合基于物品的协同过滤(item collaborative filtering,ItemCF)算法来实现短视频的推荐。在用户对某视频表现出正反馈时,使用ItemCF算法计算与该视频相似度最高且用户未观看过的N个视频,并将其插入到以LFM算法为基础的推荐列表中,结合Redis缓存数据库,实现了短视频的在线实时推荐,极大提高了用户体验,也增加了服务器的运行效率。In order to solve the problem that traditional latent factor model(LFM)algorithm cannot recommend for users in real time,on the basis of LFM algorithm,this system combines item collaborative filtering(ItemCF)algorithm to realize the short video recommendation website.When users give positive feedback to a video,ItemCF algorithm is used to calculate N videos that have the highest similarity with the video and have not been watched by the user,and insert them into the recommendation list based on LFM algorithm.In combination with Redis cache database,real-time online recommendation of short video websites is realized,which greatly improves user experience.It also increases efficiency of the server.

关 键 词：短视频网站 潜在因子模型 基于物品的协同过滤 推荐算法 实时推荐 视频推荐 相似度 用户行为 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]