Authors: Sanjeev Kumar Sharma; Ugrasen Suman
Addresses: School of Computer Science and Information Technology, Devi Ahilya University, Takshashilla Campus, Khandwa Road, Indore (M.P.), India ' School of Computer Science and Information Technology, Devi Ahilya University, Takshashilla Campus, Khandwa Road, Indore (M.P.), India
Abstract: Information overload is a significant issue of explosive growth of information on the web. The users are facing numerous problems to select and purchase interesting products online. Recommender systems are the software agents, which are helpful to reduce the problem of information overload. In this paper, architectural framework of hybrid recommender system, i.e., semantic enhanced personaliser (SEP) is proposed for web personalisation. The SEP comprised of three techniques of recommendation such as, original, semantic and category-based recommendation. The original recommendation consists of three components such as user-based collaborative filtering, item-based collaborative filtering and item-based contextual filtering. This recommendation is based on explicit feedback and contextual information provided by the web users while semantic and category-based recommendation is based on implicit feedback using data mining techniques such as, association-rule-mining (ARM), similarity measures and clustering. The SEP is capable to solve the problem of scalability, sparsity, quality of recommendation, synonymy, etc.
Keywords: data mining; recommender systems; semantic enhanced personaliser; SEP; web personalisation; recommendation systems; information overload; collaborative filtering; contextual filtering; association rules mining; ARM; similarity measures; clustering; scalability; sparsity.
International Journal of Business Information Systems, 2013 Vol.13 No.3, pp.284 - 316
Published online: 27 Sep 2013 *Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article