Data Mining Approaches in Prediction of Students' Math Achievement in PISA

• In Event: Poster Session for Quantitative Methods and Statistical Theory

Abstract

There are many studies focusing on factors affecting achievement, whereas limited research was reported regarding students’ math performance in the Programme for International Student Assessment (PISA). This study investigates the main factors that contribute to predicting students’ math achievement in PISA 2012. Four data mining approaches are implemented and compared in the data analysis including the Decision Tree, Random Forest, Artificial Neural Networks, and Support Vector Machine. The predictor variables include both the student and the school level characteristics, and the predictive accuracy is measured by several indices. As a direct outcome of this research, more efficient predictive models can be applied to improve the understanding of students’ success in math learning.

Authors

• Ya Zhang, Western Michigan University

• Abdullah Alfarwan, Western Michigan University