A Landscape of Machine Learning Models for Predicting Student Dropouts in Developing Countries: Systematic Literature Review

In Sub-Saharan countries the student dropout rate elevated to 19% of secondary school age, followed by Northern Africa and Western Asia (9%) and Southern Asia (7%) in 2019. In Tanzania, student dropouts in secondary schools increased from 3.8% in 2018 to 4.2% in 2019. The persisting dropout problem especially in secondary schools is attributed to a lack of proper identification of root causes and unavailability of formal methods that can be used to project the severity of the problem. This study applied PICO and PRISMA to conduct systematic literature reviews to analyze and synthesize and report key results. Digital databases such as ACM library, Scopus, Web of Science, and Google scholar yielded 21 exhaustive articles from January 2014 to October 2023 regarding the student dropout features identification using machine learning models. Results revealed that distance, absence, child pregnancy, early marriage, and family income are the most evident predictors for student dropout. Presented articles did not cover psychological features leading to student dropout in developing countries. Student dropouts have complex psychological and behavioral issues demand a comprehensive analysis and understanding by machine learning algorithms. Disassociating psychological factors in dropout prediction remains inappropriately answered to determine root causes of student dropouts. Moreover, this study intends to synthesize the prediction accuracy of ML models when influential features are applied to predict student dropout in developing countries. This study concluded by highlighting the most evident Machine learning algorithms that showed high prediction accuracy; Logistic Regression, Decision Tree, Random Forest, Neural Networks, and K -Nearest Neighbors.