Augmented Session Similarity Based Framework for Measuring Web User Concern from Web Server Logs

Abstract

In this paper, an augmented sessions similarity based framework is proposed to measure web user concern from web server logs. This proposed framework will consider the best usage similarity between two web sessions based on accessed page relevance and URL based syntactic structure of website within the session. The proposed framework is implemented using K-medoids clustering algorithms with independent and combined similarity measures. The clusters qualities are evaluated by measuring average intra-cluster and inter-cluster distances. The experimental results show that combined augmented session dissimilarity metric outperformed the independent augmented session dissimilarity measures in terms of cluster validity measures.