Prediction of NIRF Scores Using the Classification and Regression Tree (CART) Algorithm in Indian Engineering Institutions

Background: The National Institutional Ranking Framework (NIRF), introduced by the Ministry of Education, Government of India, in 2015, ranks higher education institutions across various disciplines, including engineering. It evaluates institutions based on five key parameters: Teaching, Learning & Resources (TLR), Research and Professional Practice (RPC), Graduation Outcomes (GO), Outreach and Inclusivity (OI), and Perception (PR). These rankings aim to promote transparency, accountability, and continuous quality improvement in the higher education sector. Methodology: A cross-sectional study was conducted using NIRF data from 799 engineering institutions in India, spanning the years 2016 to 2023. The Classification and Regression Tree (CART) algorithm was applied to predict total NIRF scores based on five predictor variables: TLR, RPC, GO, OI, and PR. Model performance was evaluated using classification metrics such as sensitivity, specificity, accuracy, Cohen’s Kappa, and the ROC curve. Feature importance analysis was conducted to identify the most influential predictors. Statistical analysis was performed using IBM SPSS 22.0. Results: The Classification and Regression Tree (CART) model predicted Total Scores primarily using RPC and TLR, generating 13 nodes with 7 terminal nodes at a depth of 3. The model showed strong performance, achieving 89% accuracy, 88% sensitivity, and 90% specificity, with a risk estimate of 19.91%. A high AUC value of 0.95 indicates a strong ability to distinguish between high and low scores. Based on variable importance, RPC was identified as the most influential predictor, followed by PR and TLR. Conclusion: The CART algorithm provides an interpretable and data-driven approach to identifying critical performance indicators among engineering institutions. By highlighting the most impactful parameters, it can support evidence-based policy decisions and institutional strategies aimed at improving national ranking outcomes.