视频节目层次化搜索和推荐方法的研究-《计算机技术与发展》

文章信息/Info

作者:: 张赟1 ; 沈兵虎1 ; 王李冬2; 1. 浙江传媒学院浙江广播电视技术研究所,杭州浙江 310018;
2. 杭州师范大学钱江学院,杭州浙江 310018

Author(s):: ZHANG Yun 1 ; SHEN Bing-hu 1 ; WANG Li-dong 2; 1. Institute of Zhejiang Radio and TV Technology,Zhejiang University of Media and Communications,Hangzhou 310018,China;
2. Qianjiang College of Hangzhou Normal University,Hangzhou 310018,China

Keywords:: hierarchical search and recommendation; heterogeneous media information data; personal preferences; graph model; program clustering

摘要:: 为高效组织视频节目并提升用户观看体验,提出了视频节目的层次化搜索和推荐方法。首先,上载本地视频节目并导入异构媒资数据,提取其中的重要信息,如标签、人物、关键帧等,并进一步采集用户个人偏好、观看记录等信息。然后,输入搜索关键词并分析其类型,确定中心人物和视频节目,实现层次化节目搜索和展示。最后,通过用户喜好、观看行为分析和节目聚类为用户推荐其可能喜欢的视频节目,并采用层次化图结构表示推荐结果。为了验证该方法的有效性,开发了软件系统,对大量用户进行了实验,结果表明,层次化视频搜索和推荐方法能够显著提升视频观看体验。

Abstract:: To efficiently organize video program and enhance users’ viewing experience,we propose a hierarchical search and recommendation method for video programs. Firstly,we upload local video program,import the heterogeneous media information data,and extract important information such as labels,characters and key frames,and then further collect users’ preferences,viewing records and other information. Secondly,we analyze the type of entered keywords,determine the central character and video program,and implement hierarchical search and display of video programs. Finally,through the users’ preferences,viewing behavior analysis and program clustering,we recommend preferred video programs for them,and display recommendation results using the hierarchical graph model. To verify the effectiveness of the method,a software system is developed and the experiment for a lot of users are carried out which shows that the pro-
posed method can significantly enhance the video viewing experience.