Explanation-Based Serendipitous Recommender System (EBSRS)

Abstract

AbstractRecommender systems (RSs) have gained immense popularity and achieved great success as intelligent information system that helps to deal with information overload problem. Recommender systems (RSs) have been very long evaluated for accuracy. Nowadays, along with the accuracy of the presented recommendation, other factors like novelty, diversity and serendipity have become an important aspect of recommendation systems. In this paper, we propose Explanation-based Serendipitous Recommender System (EBSRS) to generate explanation for the serendipitous recommendations presented to the user. Hereby, the approach integrates the concept of serendipity in recommendations, ensuring the relevance of recommendations while generating serendipitous recommendation. The approach generates the explanation for the serendipitous recommendations to provide a justification for the recommended list. The proposed approach is evaluated using accuracy and relevancy measures. Precision, recall and f-measure are used as the accuracy measure, whereas explanation coverage and unexpectedness is used to get the relevancy measure.KeywordsRecommender systemSerendipityExplanationRelevancyUnexpectedness