AI-Enhanced Counseling for Undergraduates: A Novel Application of Machine Learning for Personality Matching and Stress Assessment

This study presents a novel application of artificial intelligence techniques to address critical limitations in traditional university counseling systems. We designed, implemented, and evaluated a dual-component AI solution: (1) a rule-based personality matching algorithm for optimal counselor-student pairing and (2) a machine learning-based stress assessment system using ensemble methods. Our implementation employs a five-dimension psychological framework for personality assessment and compares three ML approaches (Random Forest, Neural Network, and Decision Tree) for stress prediction. Survey data from 98 Sri Lankan university students revealed strong acceptance of the AIenhanced approach ($80 \%$ expressed interest), with privacyconscious students showing higher acceptance rates (84.62%). Technical validation demonstrated the robustness of our approach, with the Random Forest classifier achieving 95.2% accuracy in stress prediction and identifying key predictive features (study hours, sleep patterns, and physical activity). The study establishes both technical viability and user acceptance of AI integration in mental health services for university populations, with particular emphasis on how algorithmic transparency enhances rather than diminishes user trust.