Title: Recommender using weighted frequent itemset refinement methodology of usage clusters

Authors: V.S. Dixit; Shveta Kundra Bhatia; Sarabjeet Kaur

Addresses: Department of Computer Science, Atma Ram Sanatan Dharma College, University of Delhi, New Delhi, India ' Department of Computer Science, University of Delhi, New Delhi, India ' Department of Computer Science, Indraprastha College for Women, University of Delhi, New Delhi, India

Abstract: User behaviour can be extracted from the web pages accessed by the end user while browsing the internet. This extracted behaviour helps in providing recommendations to the end user. This paper presents mining of frequent itemsets and refinement of usage clusters for web page recommendation. The not so interesting recommendations obtained from a cluster are in abundance due to large number of sessions in a cluster. To solve such a problem we intend to refine clusters on the basis of weighted frequent itemsets that in turn help to generate improved quality refined clusters. After getting refined clusters, the same can be used for a number of applications such as personalisation on the basis of interests of the end user, improvement in website structure and improving the accuracy of a recommender system. The accuracy of a centroid-based recommender system is evaluated using original and refined clusters.

Keywords: web usage mining; local frequent itemsets; LFI; global frequent itemsets; GFI; Davies-Bouldin index; Dunn's index; silhouette coefficient; recommender systems; usage clusters; recommendation systems; web users; user behaviour; internet; personalisation; website structure.

DOI: 10.1504/IJBIDM.2015.069268

International Journal of Business Intelligence and Data Mining, 2015 Vol.10 No.2, pp.95 - 122

Available online: 06 May 2015 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article