A Multi-Technique Approach for Recommender Systems
Ogunniyi T. O · 2013
Recommender Systems (RSs)are well known for their wide use in e-commerce to predict and recommend products for online users. To enhance the quality of recommendations made by such systems, different recommendation techniques have been developed. A number of hybrid approaches were proposed to minimize the limitations found in single approaches. Hybridization of Content Based Filtering (CBF) and Collaborative Filtering (CF) techniques has been used extensively for the implementation of RSs. The hybrid technique offers a high degree of effectiveness in recommendations, yet it suffers from portfolio effect characterized by data sparsity and cold start problems. Knowledge Based Filtering (KBF) are used to recommend products based on users ’ preferences, such techniques are prone to the general drawbacks of knowledge based systems. In this paper, a multi-technique approach for recommender systems is proposed. The proposed model integrates CBF, CF, and KBF approaches to give optimal recommendations to online users.