Multivariate Cluster-Ann Method For The Analysis Of Multidimensional Poverty In The Control And Public Management Of Poverty And Inequality In Colombia

Abstract

The measurement and monitoring of poverty variables are crucial for evaluating government management, serving as key indicators of the quality of life and the success of public policies. This research introduces a multivariate Cluster-ANN method to analyze multidimensional poverty, aiding in public control and management of poverty and inequality in Colombia. The methodology involves cluster analysis to identify multidimensional poverty profiles across dimensions such as education, childhood and youth conditions, employment, health, and housing in all 33 departments of Colombia. Subsequently, an Artificial Neural Network model forecasts a department's membership in a multidimensional poverty profile, revealing characteristic profiles and three poverty levels. This approach allows for the identification of departments facing critical situations. The artificial neural network model exhibits 100% accuracy in predicting department membership based on profiles identified in the cluster analysis.