Predicting ICT integration in teaching and learning in southwestern Nigeria polytechnic using ensemble-based models

This study systematically examines the factors that affect the integration of Information and Communication Technology (ICT) into educational institutions in Southwestern Nigeria, using ensemble models. ICT continues to offer important benefits for teaching and learning, even though challenges remain. We used a mixed-methods approach to identify what influences ICT adoption, build and compare ensemble-based models, and carefully assess their performance using standard classification metrics. Data were collected from structured questionnaires administered to academic staff and students at selected Southwestern Polytechnics in Nigeria. After preprocessing, the research adopt feature selection methods like Principal Component Analysis (PCA), Sequential Backward Feature Selection (SBFS), Sequential Forward Feature Selection (SFFS), and Central Tendency Feature Selection (CTFS). The data analysed with AdaBoost, Gradient Boosting Machine (GBM), Stacking, and Voting ensemble models. The Voting ensemble model performed best, reaching 98.77% accuracy and a 98.81% F1-score when Sequential Backward Feature Selection was used. These results show that combining several base learners can improve prediction performance. The study also identifies key barriers to ICT integration, helping educational stakeholders create targeted strategies to encourage ICT adoption and support inclusive, high-quality education.