Evolution of Recommender System: Conventional to Latest Techniques, Application Domains and Evaluation Metrices
Vikas Sethi, Rajneesh Kumar, Suresh Chand Gupta, Sunny Kuhar, Sumit Kumar Rana · 2024
Today’s world needs recommender systems because of rising Internet availability, rising personalization tendencies, and shifting computer usage patterns. Overloaded with information is a problem that is typically solved by recommendation systems (RS). Nowadays, there is a lot of information being produced, which makes it challenging for users to locate the pertinent information about goods and services that suit their tastes and interests. Although recommender systems have become prevalent in e-commerce, multimedia, tourism, and many other industries and are effective at producing good suggestions, they continue to face difficulties including cold start, sparsity, scalability, and many more. Scholars have gotten more and more engaged in RSs in the past few years, and numerous analyses of literature have been completed in order to look into the attributes, difficulties, and methodologies of various RSs. But neither of the preceding audits offered an in-depth investigation of every one of RSs. For the purpose of educating and directing novice researchers who are interested in the topic, the authors of this paper offer a thorough evaluation of the recommendation system. In order to act as a guide for future research and practise in this area, this paper examines the fundamental methodologies and widely used approaches, including machine and deep learning techniques in recommender systems with assessment metrices as well as the application areas of recommender systems.