Construcción y medición de un indicador socioeconómico mediante el análisis de componentes principales no lineal y técnicas de remuestreo.

Abstract

This project aims to analyze the variables of the 2010 Population and Housing Census with greater discriminating capability in order to create socioeconomic indicators at the provincial level. For this, we create a sample design in each province using stratified sampling and we apply Nonlinear Principal Components Analysis on the census data, for this allows the incorporation of qualitative and quantitative variables. Afterwards, we analyse the stability of the obtained results using the Bootstrap resampling technique, through the creation of confidence intervals. Finally, the results suggest that some census variables do not discriminate the households and that the method is stable, allowing the creation of socioeconomic indicators in each province of the country.