A Case Study on Correctness Evaluation of Content Based Recommender System Based on Text, Semantic Text and Visual Similarity

Abstract

Recommender systems provide suggestions to appropriate recipients based on suggestions and inputs given by the clients or users. The system guides the user in a tailored way to objects of interest in a larger space of possible options. In this paper, with the help of a case study the analysis of the results are obtained for different algorithms used for recommender systems—text based, semantic text based, visual similarity based product similarity. Correctness measure is performed using mean reciprocal rank metric. The paper concludes that no single algorithm is suitable for all recommender systems. Based on the type of user preferences, ensemble and hybrid algorithms are the future for recommendation.