Predicting educational attainment in Kenya: a machine learning approach using socioeconomic and geographic data

Abstract

Educational disparities in Kenya remained a critical challenge, particularly between urban and rural regions and across socio-economic groups. Despite the implementation of FPE and FSE, inequalities persisted, impacting students’ ability to complete various educational levels. This study aimed to address this issue by leveraging ML to predict educational attainment using key variables such as household wealth, access to services, and geographic data from the Kenya DHS. A ML model was developed to analyze these socio-economic and geographic factors, providing policymakers with data-driven insights to design targeted interventions aimed at closing educational gaps. By deploying the ML model through a web-based tool, stakeholders were able to identify at-risk regions and populations, leading to more effective resource allocation. The anticipated outcome was a more equitable education system, contributing to Vision 2030 by improving long-term educational outcomes and reducing socio-economic barriers to learning.