Predicting High-Ability Students by Using the Teacher Rating Scale: Insights from Machine Learning and Explainable AI Approaches

• In Event: AI, Machine Learning, and Data Science in Modeling and Prediction: New Strategies and Applications

Abstract

This study used machine learning (ML) and explainable artificial intelligence (XAI) to examine items from the HOPE Teacher Rating Scale about students' GPA. Using a sample of Syrian refugee students, the ML model achieved high predictive accuracy. The SHAP analysis revealed that demographic variables such as school level and refugee camp location status had a significant impact on the results while traditional methods identified item 1 (potential) as the primary predictor. The results from subgroup analysis showed different item importance patterns between schools and refugee camps which indicates the requirement for customized educational strategies according to student backgrounds and developmental stages.

Authors

• Taeyong Kim, Lewis University

• Hyeseong Lee, Lewis University

• Ali M. Alodat, Qatar University and Yarmouk University