A Novel Model for Detecting P2P Downloading Based on Data Mining
Wang Xiu-ying, Zhiqing Shao · Huadong Li-Gong Daxue xuebao · 2008
A novel real-time model to detect P2P downloading is presented.Firstly,features are(extracted) by using Apriori algorithm,and the dataset is offline.Secondly,P2P users are detected by clustering algorithm with the extracted features.The clustering is based on similarity among records.The(alarm) will be triggered,if the statistic value is greater than a threshold,and the dataset is online.There are three characteristics in the proposed model.The first is unsupervised approach.The second is(real-time.) The third is that the firewall is associated to cut down the connection between P2P user and(Internet.)