A Personalized Academic Advisory Recommender System (PAARS): A Case Study

Recommender systems (RSs) are effective tools to reduce the number of available selections and to expose users to items that meet their needs. RSs predict the preferences of users and generate recommendations based on the interest model of users. RSs have been used in several domains including e-learning and e-library. In particular, collaborative-based filtering (CF) and content-based filtering are the common approaches used to recommend items to users. The development of CF introduces content-based filtering which does not need users' rating for items used to compute the similarity between users or items in CF. Instead, the affinity in content-based filtering RSs is computed based on the provided information of items selected by a user and then make the recommendation accordingly. The traditional system of recommending a list of courses to students is time-consuming, risky, and monotonous work. These drawbacks may negatively affect the students' academic performance and learning experience. While content-based filtering can introduce a solution to automate the process of course selection, this paper introduces a Personalized Academic Advisory Recommender System (PAARS) that recommends a list of courses based on each student's profile and similar students' profiles. The primary data mining technique used to learn profiles for students is a k-nearest neighbors (kNN) classifier. The objectives of this research are twofold. The first is to introduce a model that personalizes the learning process, since each student may have different objectives than other students. The second is to introduce PAARS web-based framework that automates the process of course recommendation. PAARS would help students to enhance their academic performance and improve their level of loyalty to their universities as well.