万倩,谢峰,赵明,等.基于大数据的数字电视用户关机行为预测[J].电视技术,2016,40(12):68-72
基于大数据的数字电视用户关机行为预测
Prediction of shutdown behavior of digital TV users based on big data
投稿时间: 2016-05-15  最后修改时间: 2016-06-17
DOI:10.16280/j.videoe.2016.12.013
中文关键词:  大数据  关机模型  回归决策树  GBDT
英文关键词:Big data  shutdown model  regression decision tree  GBDT
基金项目:
           
作者单位
万倩 国家新闻出版广电总局广播科学研究院
谢峰 北京数码视讯科技股份有限公司
赵明 国家新闻出版广电总局广播科学研究院
王炳飞 北京数码视讯科技股份有限公司
摘要点击次数: 43
全文下载次数: 
中文摘要:
      广电有线运营商利用双向数字电视机顶盒回传的用户收视行为数据,结合广电BOSS系统提供的媒资数据,通过采用大数据分析技术,为数字电视用户提供个性化的收视服务,极大提升了用户体验并增加了营业收入。然而,回传数据难以捕捉到电视机的状态,在机顶盒未关闭而电视机关闭的情况下,通过回传数据得到的用户收视行为是无效的,会影响大数据分析系统对用户收视行为的预估。本文通过分析已知的数字电视用户关闭电视机的行为特征,生成用户关机模型,从而预测无法采集到电视机关机数据的数字电视用户的关机行为,保证用户收视行为统计的有效性。
英文摘要:
      Radio and television cable operators greatly enhance the user experience and increase operating revenue by providing personalized viewing service for digital TV users, which benifits from the big data techniques and the abundant data collected by two-way digital TV set-top boxes and TV boss system. However, it is difficult to collect TV state for the set-top boxes except rare TVs Therefore, when the set-top box is not closed and the TV is turned off, the user viewing behavior uploaded during this period is invalid. which will affect the big data analysis system’s prediction of the user viewing behavior.With the deep analysis of TV shutdown behavior on partial users, this paper builded an user shutdown model, to predict the shutdown behavior of other digital TV users whose shutdown data cannot be collected. Consequtently, the validity of user viewing behavior statistics is ensured with this model.
查看全文  查看/发表评论  下载PDF阅读器
关闭