无监督的视频业务特征分析与分类  

Unsupervised Feature Analysis and Classification of Video Streams

作　　者：姚利涛 董育宁[1] 

机构地区：[1]南京邮电大学通信与信息工程学院,南京210003

出　　处：《应用科学学报》2015年第2期117-128,共12页Journal of Applied Sciences

基　　金：国家自然科学基金(No.61271233;No.60972038);教育部博士点基金(No.20103223110001)资助

摘　　要：基于机器学习的流统计特征识别的方法关键在于如何找到具有区分力度的业务流统计特征.为此,提出了一些能够较好地区分视频业务的QoS相关的统计特征.为了充分地发挥多级聚类算法的优势,以灵活的特征选择策略标记不同层级的网络视频流,通过大量的真实网络视频数据进行实验验证.结果表明,该方法能比现有同类方法取得更高的分类准确率.For recognition of flow statistical features based on machine learning, the key is to select distinguishable features of different traffic flows. This paper presents several QoS- related statistical features that can well discriminate video traffics. To make full use of the advantages of hierarchical clustering algorithm, this paper uses a flexible feature selection strategy to mark the network video streaming of different levels. Experiments are performed on a large scale real network video data. The results show that the proposed method can achieve significantly higher classification accuracy compared to existing methods.

关 键 词：视频流 统计特征 QOS 流分类 多级聚类 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]