The State of the Art in Deep Learning-based Recommender Systems

Abstract

AbstractRecommender systems are tools and techniques for effective information filtering that have become more common as a result of increased Internet access, personalization tendencies, and changing computer user habits. Existing recommender systems are capable of making acceptable recommendations, but they face issues in recommending accurate items, suffer scalability problem with big data and cold start. In recommendation systems, deep learning, a recently created sub-domain of machine learning technique, is used to improve the feature of anticipated output. Deep learning is utilized to generate suggestions, and the research issues that come with employing deep learning for recommendation systems are also discussed. The core terminology, fundamental principles of a recommendation engine, and a comprehensive overview of deep learning models used in recommender systems are discussed in this paper.KeywordsRecommendation systemCollaborative filteringDeep learningAutoencodersNeural networks