Aplicación de modelos no paramétricos para predecir y explicar el riesgo de quiebra empresarial del sector industrial manufacturero ecuatoriano.

Abstract

The objective of this paper is to apply non-parametric methods to predict the risk of Ecuadorian manufacturing business bankruptcy in the period 2016-2020, analyzing the financial behavior involved in a manufacturing company. In the first instance, a study analysis is carried out on the manufacturing industry, in order to analyze the influence of this economic sector on the economic development of the country. In addition, we proceed to study non-parametric such as the generalized additive model (GAM) and the artificial neural network (RNA) technique, which will be the necessary tools to estimate the calculation of bankruptcy. To obtain this result, financial indicators that best describe a bankruptcy process were previously identified through the GAM, since this technique yields more flexible and easily interpretable results. Based on the latter, it was sought to systematize the identification of bankruptcy risk and its possible causes, for which the RNA technique was applied. With this, it was possible to predict the manufacturing companies that entered bankruptcy proceedings based on the financial information identified by the GAM. Thus, it was possible to apply both models, in such a way that one complements the other based on the interpretation.