Modelamiento por componentes principales con el enfoque del análisis de datos funcionales para predecir el riesgo país del Ecuador a través del Indicador de Bonos de Mercados Emergentes.

Abstract

This project investigates the application of Functional Data Analysis (FDA), a relatively new technique in the field of Statistics, addressing from initial concepts such as: functional random variable, functional data, adjustment of discrete data from smooth functions and statistical measures of centrality and variability, to then study the functional approaches of depth measurements, adopting the concept of the most central curve, and the method of functional principal components, useful for the elaboration of a model which predicts a complete function. This technique is applied to the daily data of the Ecuadorian country risk to obtain the future behavior of the daily scores for the year 2021; country risk is a benchmark for attracting investors and contributing to the development of a country, determined by the Emerging Markets Bond Index (EMBI). In addition, for the model, two forecasting scenarios are proposed, based on the best and worst case and, thus, understanding the future trend of the indicator in both scenarios and providing solutions to the most serious impacts that this indicator may cause