Peer-to-Peer Network Classification Using nu-Maximal Margin Spherical Structured Multiclass Support Vector Machine

Abstract

In this paper we propose a P2P network traffic classification method using nu-Maximal Margin Spherical Structured Multiclass Support Vector Machine (nu-MSMSVM) classifier. The P2P network traffic is classified into different classes based on four applications namely, Bit Torrent, PPLive, Skype and MSN. The concept of Hypersphere based classifiers being able to minimize the effect of outliers has been adapted in this work. The experimental results show low false positive and false negative ratio thereby achieving high precision and recall rate.