摘要
为了进一步提高基于P2P的内容分发网CDN(content delivery network)的搜索效率,对用户就近组成的节点群(peer group)进行了研究,提出了节点群相似的概念和基于关键词的节点群相似性度量模型.当一个需求在本节点群不能满足时,优先到相似性高的节点群中查找,以较快地满足跨节点群的需求.度量模型对任意两个节点群首先根据节点中各文档关键词的一致程度判断对应两个节点的相似度,然后根据相似节点对的数量和相似度来判断两个节点群相似度.实验表明,使用所提出的度量模型得出的计算结果比传统的基于VSM的算法更接近于实际情况.
To improve search efficiency in Content Delivery Network based on peer-to-peer,users weredivided into peer groups.The concept of peer group was discussed and a keyword-based method of measur-ing similarity between peer groups was proposed.When none of peers could provide the results for a re-quest,another peer group having higher similarity degree with it would be searched to meet the requestsoon.In the model,similarity degree of each peer in the two peer groups was first calculated according tothe keywords in each document,and then the similarity degree between the two peer groups was deter-mined based on the amount of the similar pairs and similarity degree.Experiments demonstrated that theproposed model got closer results with the actual situation than traditional SVM model.
出处
《武汉大学学报（理学版）》
CAS
CSCD
北大核心
2011年第6期489-493,共5页
Journal of Wuhan University:Natural Science Edition
基金
国家自然科学基金资助项目(60872051)
关键词
节点群
关键词
内容分发
相似性
相似性度量
peer group
keyword
content delivery
similarity
similarity measurement
