State of the Art: Artificial Intelligence Algorithms for Measuring Poverty Indicators

Abstract

The purpose of this work is to present a review of the state of the art regarding the most widely used algorithms for measuring poverty globally, in order to identify current analytical approaches and offer strategies that contribute to addressing this complex and dynamic problem. The review reveals that the algorithms most prevalent in recent literature are Convolutional Neural Networks (CNNs), Random Forest (RF), Gradient Boosting (including XGBoost), and Non-Convolutional Neural Networks. Convolutional Neural Networks stand out for their high effectiveness in processing and extracting visual features, while Random Forest and Gradient Boosting are notable for their versatility, predictive capacity, and robustness in diverse contexts.