Using Machine Learning to Cluster and Predict Students' Learning Styles

It is crucial to understand individual learning styles when designing personalized and effective educational experiences. This research applies machine learning techniques to predict high school students' learning styles. With data collected via a structured questionnaire, the elbow method and the k-prototype algorithm were used to identify three optimum numbers of clusters: the pragmatic/goal-oriented learners, the reflective/resilient learners, and the meticulous/methodical learners. A four-cluster configuration was also examined according to the four dimensions in the Felder-Silverman Learning Style Model. Decision Tree, Random Forest, Support Vector Machine, Logistic Regression, and Naive Bayes were employed to validate the clustering results. Random Forest and SVM recorded the maximum performance rates at 95% and 90% for the three-cluster and four-cluster configurations respectively. The results suggest incorporating machine learning in educational systems to encourage adaptive and inclusive learning environments and highlight the implication of predictive modeling in enhancing student engagement and academic results.