Analyzing Educational Disparities Between Urban and Rural Areas in Malaysia

Abstract

Educational disparities between urban and rural areas remain a critical challenge for Malaysia’s national development. This study investigates the factors driving rural–urban academic achievement gaps by combining non-parametric inferential statistics with supervised machine learning data-mining techniques. Primary data were collected from  respondents via a bilingual online survey capturing demographic attributes, learning environments, support systems, and material/ICT resource availability. Non-parametric hypothesis testing confirms a statistically significant difference in overall satisfaction with educational opportunities between independent urban and rural cohorts. To explain this divide, three machine learning models (J48, Random Forest, and Logistic Regression) were evaluated using 10-fold cross-validation in Weka to predict tertiary academic attainment. Random Forest achieved the highest classification performance (  accuracy, , ). Information Gain feature evaluation reveals that age (  bits) and family academic guidance (  bits) are the dominant drivers of higher education attainment, whereas geographical location alone contributes minimally (  bits). Furthermore, J48 decision tree logic indicates that for young rural learners with adequate home environments, access to ICT and study materials is the decisive factor in reaching tertiary education. These findings demonstrate that Malaysia's educational gap is primarily mediated by household resources and academic support rather than geography itself, suggesting that policy interventions should prioritise household-level digital access and family engagement programs over purely regional funding.