International Journal of Information Technology and Applied Sciences ISSN (2709-2208) Int. J. Inf. Tec. App. Sci. 3, No.3 (July-2021) http://woasjournals.com/index.php/ijitas 161 https://doi.org/10.52502/ijitas.v3i3.128 A Hybrid Movie Recommender System and Rating Prediction Model Muhammad Sanwal 1 , Cafer ÇALIŞKAN 2 1 Electrical and Computer Engineering, Antalya, Turkey 2 Electrical and Computer Engineering, Antalya, Turkey Abstract: In the current era, a rapid increase in data volume produces redundant information on the internet. This predicts the appropriate items for users a great challenge in information systems. As a result, recommender systems have emerged in this decade to resolve such problems. Various e-commerce platforms such as Amazon and Netflix prefer using some decent systems to recommend their items to users. In literature, multiple methods such as matrix factorization and collaborative filtering exist and have been implemented for a long time, however recent studies show that some other approaches, especially using artificial neural networks, have promising improvements in this area of research. In this research, we propose a new hybrid recommender system that results in better performance. In the proposed system, the users are divided into two main categories, namely average users, and non-average users. Then, various machine learning and deep learning methods are applied within these categories to achieve better results. Some methods such as decision trees, support vector regression, and random forest are applied to the average users. On the other side, matrix factorization, collaborative filtering, and some deep learning methods are implemented for non-average users. This approach achieves better compared to the traditional methods. Keywords: Recommender systems, matrix factorization, collaborative filtering, hybrid systems, decision tree method, support vector regression, random forest method. 1. INTRODUCTION In the current era of modern technology, the amount of information is increasing rapidly. Available information on the internet is not relevant to the users’ needs and preferences [1]. Most of the users spend their precious time navigating towards useful information. Therefore, recommender systems are getting popular especially with the rapid growth of e-commerce. Such systems provide the best solutions for this problem. E-commerce platforms such as Netflix, Amazon Prime have millions of users with millions of items to offer [2]. As a result, it is a great challenge for these companies to recommend items to the users according to their preferences. In this sense, a recommender system is one of the modern tools to solve this sort of problem in the current era. Recommender systems are categorized into three types, namely content-based filtering, collaborative filtering, and knowledge-based filtering [3]. Generally, items are recommended on a similarity basis either on a user profile or an item profile. These approaches find the similarities among users or items, then make suggestions to specific users according to their profiles. They mostly rely on explicit feedback, meaning that users provide explicit input regarding their interests in various products into the system. For example, Netflix benefits from the star rating system in which users submit their evaluations after watching movies. In the current decade, hybrid recommender systems are emerging as successful solutions as hybrid systems achieve better results compared to the conventional methods[4]. These methods have overcome the issues due to the weaknesses of recommendation techniques by replacing them with the strength of other techniques, so their performances depend upon the integration of their components. Mostly in the application, databases have large numbers of items, and this makes it very difficult for any user to view or rate all the items. As a result, every user visits or rates just a limited number of items in the related database which results in sparse user-item matrices for recommender systems. This also makes it very challenging to recommend the desired items to a specific user. Moreover, another main challenge in this field is fetching the appropriate features for the items themselves.