ABSTRACT

Clustering web sessions is to group web sessions based on similarity and consists of minimizing the intra-group similarity and maximizing the inter-group similarity. Here in this paper we developed a new similarity measure named SSM(Sequence Similarity Measure) and enhanced an existing DBSCAN and OPTICS clustering techniques namely SSM-DBSCAN, and SSM-OPTICS for clustering web sessions for web personalization. Then we adopted various similarity measures like Euclidean distance, Jaccard, Cosine and Fuzzy similarity measures to measure the similarity of web sessions using sequence alignment to determine learning behaviors of web usage data. This new measure has significant results when comparing similarities between web sessions with other previous measures. We performed a variety of experiments in the context of density based clustering, using existing DBSCAN and OPTICS and developed SSM-DBSCAN and SSM-OPTICS based on sequence alignment to measure similarities between web sessions where sessions are chronologically ordered sequences of page visits. Finally the time and the memory required to perform clustering using SSM is less when compared to other similarity measures.