Exploiting swarm behaviour of simple agents for clustering web users' session data

Abstract

In recent years the integration and interaction of data mining and multi agent system (MAS) has become a popular approach for tackling the problem of distributed data mining. The use of intelligent optimization techniques in the form of MAS has been demonstrated to be beneficial for the performance of complex, real time, and costly data mining processes. Web session clustering, a sub domain of Web mining is one such problem, tackling the information comprehension problem of the exponentially growing World Wide Web (WWW) by grouping usage sessions on the basis of some similarity measure. In this chapter we present a novel web session clustering approach based on swarm intelligence (SI), a simple agent oriented approach based on communication and cooperation between agents. SI exploits the collective behaviour of simple agents, cooperation between the agents, and emergence on a feasible solution on the basis of their social and cognitive learning capabilities exhibited in the form of MAS. We describe the technique for web session clustering and demonstrate that our approach perform well against benchmark clustering techniques on benchmark session data. © 2009 Springer-Verlag US.