Desarrollo de un sistema de alerta temprana sobre riesgo institucional financiero: estudio de casos en Ecuador.

Abstract

This Degree Project proposes to establish a banking risk rating system at the national level that allows forecasting the state of banking fragility in which the entities will be during future periods, this based on both micro and macroeconomic indices that will be the financial signals of alert to be used by each bank to predict a possible financial destabilization. Within the study carried out, the theory and methodology of Machine Learning and Hidden Markov Models (HMM) were applied with the objective of decoding the risk levels that financial entities will have in a certain period based on a probability initial transition from one state to another with the same probability. This was done by using the statistical software R-STUDIO.