基于Logistic模型的P2P资源流行程度分析  

Logistic-based Analysis for P2P Resources Population

作　　者：张涛[1] 张建标[1] 林莉[1] 

机构地区：[1]北京工业大学计算机学院,北京100124

出　　处：《北京工业大学学报》2013年第3期407-411,429,共6页Journal of Beijing University of Technology

基　　金：国家"973"计划资助项目(2007CB311100);北京市自然科学基金面上资助项目(4122012);北京工业大学博士启动基金资助项目(00700054R1764)

摘　　要：为了解决目标资源数量增多而带来的监管成本过高的问题,提出一种基于Logistic模型的资源流行程度分析方法.利用该方法估计目标资源的流行程度,剔除流行程度低的资源,以此缩小监控范围,减少节点的部署数量.在Kad中对不同资源进行追踪,统计资源的共享节点数量,估计Logistic模型的参数,并进行曲线拟合.通过决策树建模得出,资源的共享节点数量和资源大小是影响该方法准确性的关键因素.To solve censor cost problems brought by targets increasing, an logistic-based approach which was taken to analyze population of P2P resources was proposed. Nodes scale would be controlled by excluding targets with low population estimation. For a number of resources in Kad, the statistics of downloading nodes were made, the parameters were estimated and the logistic model was fitted with the collected dataset. By decision tree modeling, conclusion shows that population and file size are key factors to logistic-based analysis.

关 键 词：对等网络 Logistics模型 流行程度统计 决策树 

分 类 号：TP393.0[自动化与计算机技术—计算机应用技术]