Mining of Classification Trees to Analyze a Multidimensional Phenomenon

Abstract

During periods of remarkable trade openness, increase income inequality in many countries. This paper analyzes how factors that influence inequality due to commercial globalization interact each other. For which a reliable Classifier Tree-selected through a modeling process of bootstrapping-is built, it has 14 knowledge rules and classifies 84% of the observations correctly. This model indicates that inequality's changes into a country, due greater economic integration, depend principally on the labor market' structure-in agricultural countries and urbanization processes (industrialization) it reduces depending in turn on the rule of law; on the other hand, in countries with a strong service sector and good trade terms it increases in periods of stagnation or with low levels of high technology exports.