A Machine-Learning Approach to Predicting the Achievement of Australian Students Using School Climate; Learner Characteristics; and Economic, Social, and Cultural Status

Datos Bibliográficos

ID22044779
AutoresMyint Swe Khine (0000-0003-2582-7306, Curtin University, autor de correspondencia), Yang Liu (0000-0003-1220-7044, Shanghai Maritime University), Liu Yang (0000-0003-3048-3392, Shanghai Maritime University), Vivek K Pallipuram (0000-0002-3721-3814, University of the Pacific), Ernest Afari (0000-0003-2814-4439, University of Bahrain)
Año2024
Volumen14
Número12
Páginas1350
Fecha de publicación2024-12-10
Peer ReviewedSí
Open AccessSí
TipoARTICLE
RevistaEducation Sciences (JOURNAL)
Identificadores de la revistaISSN: 2227-7102 • E-ISSN: 2227-7102
EditorialMDPI AG (PUBLISHER • IT)
DOI10.3390/educsci14121350
OpenAlexW4405240137
IdiomaEN
Referencias citadas70

The Programme for International Student Assessment (PISA) is a global survey conducted by the Organisation for Economic Co-operation and Development (OECD) to assess educational systems by evaluating the academic performance of 15-year-old school students in mathematics, science, and reading. In PISA 2022, 13,437 students from Australia participated in the test. While the PISA main questionnaire assesses the subject knowledge, the student background questionnaire solicits contextual information such as school climate, learner characteristics, and socioeconomic status. This study analyses how these contextual variables predict student achievement using the machine-learning models Ridge Linear Regression, K-Nearest Neighbours, Decision Trees, eXtreme Gradient Boosting, and Support Vector Machines, and it reports the evaluation matrices and the most accurate model in predicting student achievement. The analysis shows that contextual variables are associated with student achievement and account for 42% of the variance in achievement. In addition to evaluating multiple machine-learning regressors, Shapley Additive Explanation (SHAP) analysis is conducted to explain the model predictions and evaluate feature importance. Using SHAP analysis, this paper demonstrates how educators and school administrators may effectively interpret the machine-learning results and devise strategies for student success

Academic achievement · Mathematics education · Computer Science · Early Childhood Education and Development · Education Methods and Practices · Educational Environments and Student Outcomes · Psychology