Abstract
Recommender systems aim to deliver users with intelligent methods for navigation and identification of complex information spaces, especially in the e-commerce realm. However, these systems need to overcome certain limitations that adversely impact their performances, such as overspecialization of recommendations and cold-start problem. To address these concerns, we propose a case-based recommendation approach, which is a form of content-based recommendation, in this paper. This approach is well-suited to many product recommendation domains owing to its clear organization of users' needs and preferences. Additionally, it employs a feature weighting technique to improve the performance of the recommender by enhancing its accuracy and precision. We herein present the results for different case structures, different numbers of similar cases retrieved, and various feature-weighting approaches, which indicate that better results are obtained with the proposed recommender when compared to the KNN retrieval algorithm.