SDCS-ALBPNN model for data-driven teaching quality evaluation in higher education

Abstract Evaluating teaching quality in higher education remains a critical but complex task due to the prevalence of subjective assessment methods, inconsistent data standards, and limited real-time feedback mechanisms. To overcome these challenges, this research introduces an advanced Neural Network model integrated into a Comprehensive Education Management Platform (CEMP) for objective, data-driven evaluation of teaching effectiveness. The proposed model leverages multi-source data, including student academic performance, teacher evaluation scores, attendance records, and student feedback, to provide a holistic and evidence-based assessment framework. Data preprocessing involved handling missing values and normalizing numerical features using Z-score normalization to ensure consistent input scaling. A Snap-Drift Cuckoo Search-Driven Adaptive Learning Backpropagation Neural Network (SDCS-ALBPNN) is employed for evaluation, given its strong ability to capture nonlinear relationships in educational data. The model classifies teaching quality into four categories: Excellent, Good, Average, and Poor. Experimental results demonstrate that the proposed model achieves high classification performance, with notable precision (95.2%), accuracy (98.4%), F1-score (96.1%), and recall (97.3%) across all categories. These results indicate improved classification performance under the evaluated experimental conditions. This research contributes a scalable, intelligent evaluation framework that has the potential to support continuous monitoring and feedback within educational institutions. By integrating artificial intelligence into academic decision-making, the proposed approach has great potential to improve data-informed learning enhancements and quality assurance processes of contemporary higher education.