Design and Implementation of Architectural Framework of Recommender System for e-Commerce
Sanjeev Kumar Sharma, Devi Ahilya, Ugrasen Suman · 2011
The rapid development of internet technologies in recent decades has imposed a heavy information burden on users. The popularity of recommender systems has evolved to provide suggestions and recommendations to the user for relevant information from web according to their preferences. Most recommender systems use collaborative-filtering or contentbased methods to predict new items of interest for users. While both methods have their own advantages, individually they fail to provide good recommendations in many situations. In this paper, we propose a standard architectural framework Semantic Enhanced Personalizer (SEP) which integrates three recommendation techniques i.e., original, semantic and categorybased. This framework fulfills user-based and item-based approach of recommendations. The original recommendation will be based on contextual information and the ratings provided by users explicitly, while, the semantic and category-based recommendation will be based on various data mining techniques such as, association rule mining, clustering and similarity measures. This framework overcomes the problem of cold-start