Predicción del gasto público en los Municipios de la Región Andina de Venezuela empleando Redes Neuronales Artificiales

In the budget law project prepared annually in Venezuela there are considered the objectives and goals of the National Strategic Plan and the Annual Operational Plan (POA), the estimation of public expenditure is made considering as key variables the price of oil and an exchange rate estimated for next year. By not considering external variables such as inflation, increases in wages and salaries, among others, there is a budget deficit that must be corrected through additional credits. This research presents an application of artificial neural networks for the prediction of the expenditure of the municipal entities of the Andean Region of Venezuela, incorporating information on socioeconomic and demographic variables. An exhaustive univariate and multivariate statistical analysis is performed, models are adjusted to perform two types of prediction: individual prediction for each municipality and average prediction for the groups identified in the statistical analysis.