摘要
目前,视频服务已经成为用户浏览互联网时最关注的服务之一。因而,提升在线视频服务的用户体验具有重要的研究价值。为此我们设计并实现了一个在线视频服务测量系统,利用部署在不同位置、运营商网络的测量节点设备测量在线视频服务的性能,根据测量结果对视频服务网站的特性进行分析,进而提出优化建议。在本文中,我们提出了一种利用信息增益的视频服务性能分析方法,基于此对视频服务测量系统采集的一个月的数据进行了深入分析和特征提取,验证了不同元数据和服务质量之间的关系。研究发现,服务质量和测量节点的网络位置的关系最为显著,其他相关性较为显著的元数据包括测量时间、视频分类和视频热度。
Nowadays,video service have been one of the most interest service when user surf the Internet. Therefore,it's has significant research value to raise the user's experience of service. To this end,we proposed and implemented a Measuring System for Online Video Service,which can measure the quality metrics of online video by measurement nodes in different locations and different ISPs.Also the system can analysis the characteristic of online video service website and propose suggestion to optimization. In this paper,we perpose an analysis method based on Imformation Gain. With a dataset of one month data collected by the system,we analyzed deeply and did Feature Extraction,and verify the relationship between metadata and quality metrics. This work shows that the quality metrics have the strongest correlation with the location of measurement nodes. Other online video metadata as measure time,type of video and heat of video also play a role with quality metrics.
作者
卓煜
尤佳莉
王劲林
ZHUO Yu YOU Jiali WANG Jinlin(National Network New Media Engineering Research Center, Institute of Acoustics Chinese Academy of Sciences, Beijing 100190, China University of Chinese Academy of Sciences, Beijing 100190, China)
出处
《网络新媒体技术》
2017年第2期1-6,共6页
Network New Media Technology
基金
智能电视平台与服务支撑环境研制(XDA06040501)
中国科学院青年创新促进会(Y529111601)