Applied Mathematics & Information Sciences

Author Country (or Countries)



Over the past decade, Machine Learning has become a practical approach for simulating and examining social issues, notably poverty, education, and health diseases. This study compares the performance of various machine learning methods especially Support Vector Machines (SVM), Decision Trees (DT), and Logistic Regression (LR) in predicting poverty status. For this purpose, the present contribution employs a micro dataset which has been extracted from the National Survey on Household Consumption and Expenditure 2013/2014. Several evaluation metrics such as accuracy, precision, Cohen’s Kappa statistic, F1-score, and recall are used to evaluate the models’ outputs. The R results indicate that the three algorithms achieved high accuracy scores. Therefore, the decision trees have more improvements in terms of accuracy (99.61%) compared to LR (91.09%) and linear kernel SVM methods ( 99.24%).

Digital Object Identifier (DOI)