CAREERNET: A HYBRID DEEP LEARNING AND REINFORCEMENT LEARNING MODEL FOR STUDENT CAREER RECOMMENDATIONS

Abstract

The increasing complexity of career decision-making among undergraduate students highlights the need for intelligent, personalized, and sustainable guidance systems. This paper introduces CareerNet, a hybrid framework that integrates deep learning with reinforcement learning to provide tailored career recommendations while accounting for both skill competency and mental well-being. Leveraging the OCEAN personality model and aptitude assessments (numerical, spatial, perceptual, abstract, and verbal) from a publicly available Kaggle dataset comprising 2,000 instances, the framework constructs comprehensive student profiles that capture cognitive strengths, behavioral traits, and mental health indicators. The proposed CARE-RL (Career and Resilience-Enhanced Reinforcement Learning) algorithm models career selection as a multi-objective optimization problem, where the reward function balances skill-career alignment with mental health sustainability. CareerNet employs deep feature embeddings for student profiling, graph-based career mapping, and adaptive decision-making through reinforcement learning, enabling dynamic and explainable recommendations. Experimental results demonstrate superior performance compared to traditional classifiers and collaborative filtering approaches, achieving improvements in accuracy, top-N recommendation precision, and interpretability. This research contributes a scalable, explainable, and student-centric AI framework that bridges the gap between academic skills, mental health, and career development, offering actionable insights for both students and educators.