Abstract
A large-scale regression experiment to analyze the comprehensive determinants of income inequality across many countries has been carried out. A support vector machine (SVM) method was applied to the nonlinear regression analysis of the relation between Gini coefficients (as a dependent variable) of 161 countries and 57 factors (as explanatory variables) in economic, political, educational, health and technology fields. The optimum determinants for the income inequality were searched by using the sensitivity analysis method as a variable selection technique in SVM. It is found that 25 factors satisfactorily reproduce the Gini coefficients of 161 countries with a high coefficient of determination (R2) of 0.795. It is revealed that, among 25 determinants, political factors contribute most significantly to the income inequality, while economic factors are next important, and educational factors little contribute to inequality.
