A Comparative Overview of Hybrid Recommender Systems: Review, Challenges, and Prospects
Rakhi Seth, Aakanksha Sharaff · 2022
Recommender System (RS) helps to find the items according to user interest and provides various suggestions that help in the decision-making process. These suggestions depend on distinct recommendation techniques. These approaches are divided into different categories like Collaborative, Content, Demographic, Utility, Knowledge-based, and Hybrid. Collaborative RS works on the concept of “people to people co-relation”. Content-Based RS suggests the idea of recommendation in which next item for a user is similar to the item that user like in the past. Demographic RS categorized the attributes based on the demographics of the user or item and make recommendations based on these demographics classes. Utility-based RS is a concept which depends on the estimation of the utility for a user for each item by using a specific utility function. Knowledge-Based RS works on “a particular user needs that how it meets with the item”. Hybrid RS is a combination of two or more recommendation strategies in a distinguished way to create a better and more personalized experience for the individual user like one of the recent study suggested the novel approach of solving sentimental issues faced in recommendation system by combining the two models of collaborative filtering i.e., by taking the rating score from memory-based and tagging (vital role in describing user feelings using Matrix Factorization (MF)) approach from model-based. So our objective is to find out how the hybrid approach is better and also understand the different types of solutions using the hybrid recommender systems and demonstrating new challenges and future scope of the hybrid system. The basic and widely discussed problem of the traditional RS technique is cold start problem and the problems associated with demographics, the utility is also considered for the study and how the Hybrid based approach is used to analyze and solve these problems.