Indian Journal of Science and Technology

Abstract

Objectives: Today the recommendation technology has managed to achieve a distinct place in the modern and fascinating world of e-commerce applications as it helps the user in selecting items or products of his interest from a large pool. The present article aims to provide a comprehensive and systematic review of the state-of-the-art recommender systems. Methods/Statistical Analysis: The entire literature review process was divided into six research questions keeping in view the different perspectives of recommendation field. The methodology adopted here, consists of the search plan and the paper selection criteria. The search plan attempts to retrieve the research studies through several digital libraries and the paper selection criteria help filter out the most relevant studies further to gather evidence against each of the research questions. Findings: The literature review process provides a thorough discussion on different techniques deployed in recommender system literature such as collaborative filtering, content-based filtering, social filtering, demographic and knowledge-based and utility based systems. It also explores their strengths and weaknesses. The recommender systems face certain challenges in their deployment such as cold-start, sparsity, scalability, user privacy, etc. The different application domains where recommender systems are being adopted these days include movie, music, books, news, tourism etc. The gap analysis conducted during literature review, focuses on improving the traditional recommendation approaches, the precise blend of existing approaches with different types of information, modeling of user profiles and recommended items, standardization of non-standard evaluation techniques etc. Application/Improvements: This paper also throws some light on certain application fields such as television, research grants, restaurant, job search, etc. that need to grab the attention of scientific and research communities to promote more research in those areas.

Keywords: Application Domains, Collaborative Filtering, Evaluation Measures, Recommendation Approaches, Recommender Systems