Quantum -Enhanced Recommender System for Personalized Elective Selection in Education

Abstract

Elective courses selection is one of the most important decisions undertaken by undergraduate students as it directly affects one's academic success and career path. Yet, the course selection is often based on advice from friends, family, or faculty, leading students to select courses incongruously based on their academic aptitude. If we could develop a personalized elective recommendation framework, then students can select their electives accordingly based on their ability. In this paper, a Hybrid Quantum-Classical Elective Recommendation System (HQ-ECRS) is presented that integrates the nine-qubit Variational Quantum Classifier (VQC) with Logistic Regression to formulate a personalized recommended elective courses framework. In our study, a statistically representative synthetic undergraduate dataset comprising 10,000 records of students with nine college attributes including CGPA (range 0-10), Average Exam Score, Mathematics Score, Programming Score, Data Structures Score, Algorithms Score, DBMS Score, Statistics Score, and Electronics Score were used to develop and evaluate our framework, which generates a ranking of nine types of Elective category and recommends the top 3 suitable Elective Course Categories to students accordingly. Our VQC utilizes data re-uploading, angle embedding, and strongly entangling layers for nonlinear modelling and Logistic Regression for stable probability estimates, then fuses the nodes using weighted hybrid fusion. With Flask as the web application framework as a demo prototype, the framework support login function, recommendation history, and visualization of prediction result. When compared against the four classical algorithms, Logistic Regression, Decision Tree, Random Forest, and K-Nearest Neighbor's, our hybrid quantum-classical framework yielded a good accuracy on undergraduate elective courses selection forecast, and successfully outperformed all classical baselines.