基于用户偏好挖掘算法的IPTV用户多维画像的设计与实现  

Design and Implementation of IPTV User Multidimensional Profile Based on User Preference Mining Algorithm

作　　者：张智骞 丁凤 郭永康 宋世聪 Zhang Zhiqian;Ding Feng;Guo Yongkang;Song Shicong(Guangdong South New Media Co.,Ltd.,Guangdong 510012,China;School of Information and Communication Engineering,Communication University of China,Beijing 100045,China)

机构地区：[1]广东南方新媒体股份有限公司,广东510012 [2]中国传媒大学信息与通信工程学院,北京100045

出　　处：《广播与电视技术》2024年第8期59-63,共5页Radio & TV Broadcast Engineering

基　　金：企业委托合作项目“基于媒体融合与传播的元宇宙及大数据技术专项课题研究项目”(No.HG23002)资助

摘　　要：随着用户增长和视听类业务发展,IPTV平台积累了海量的用户行为数据,精准挖掘行为数据中的用户偏好是视听类业务持续发展的关键。本文针对IPTV用户特点和业务场景需求,基于IPTV用户行为数据设计了统计规则类和机器学习类的用户偏好挖掘算法,分别挖掘IPTV用户的基础偏好和深层次偏好,并通过Hadoop、Hive、Spark等大数据技术实现了基础标签和深层次标签相结合的IPTV用户多维画像。本文的研究可为网络视听媒体资源价值管理、定制化服务和个性化推荐提供依据。With the growth of users and the development of audio-visual services,IPTV platforms have accumulated massive user behavior data.Accurately mining user preferences in behavioral data is the key to the sustainable development of audio-visual services.According to the characteristics and business scenarios of IPTV users,this paper designs a user preference mining algorithm based on statistical rules and machine learning for IPTV user behavior data,respectively mining the basic and deep preferences of IPTV users.It realizes the combination of basic and deep labels of IPTV user multidimensional portrait through big data technologies such as Hadoop,Hive,and Spark.The research in this paper can provide a basis for media resource value management,customized services,and personalized recommendations.

关 键 词：IPTV 用户行为 用户偏好挖掘 用户多维画像 标签 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]