A Hybrid Recommendation System for University Selection Using Machine Learning

Research Problem

University selection requires balancing academic eligibility, costs, program relevance, and personal preferences. This research explores how quantitative data and qualitative information can be combined to make university comparisons more personalized and useful.

Methodology

Match-Score Prediction: XGBoost regression predicts university match scores using acceptance rates, tuition fees, GPA, and GRE scores.

Course Relevance: Cosine similarity compares unstructured course descriptions with student interests.

Reputation Analysis: Sentiment analysis evaluates university reviews as an additional input to decision-making.

Application Assistance: The proposed system includes deadline monitoring, side-by-side university comparison, and template-based NLP for personalized Statements of Purpose.

Key Findings

The paper presents a hybrid approach that combines quantitative and qualitative information for personalized university recommendations. The authors report testing the system’s operational efficacy, without claiming specific accuracy scores, dataset sizes, numerical improvements, or comparative performance results.

Publication Details