基于Logistic模型的P2P资源流行程度分析

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.