Modelos estadísticos para la detección de patrones en Medio Ambiente y Finanzas.

Abstract

Environmental pollution from greenhouse gas (GHG) emissions is con sidered a challenge for many researchers. Among the greenhouse ga ses, carbon dioxide (CO2) has generated massive effects on air quality in large cities and is considered as an alarming problem worldwide sin ce CO2 emissions also affect human health, natural life and agriculture. The main objective of this research work is to study and estimate CO2 emissions from different sources of energy consumption in Ecuador. A data set collected during the years 1965 and 2020 was used. The esti mation and forecasting of CO2 emissions is performed by developing a Multiple Linear Regression Model (MLR) using the least squares method, a Generalized Potential Regression Model (GPR) using the Particle Swarm Optimization (PSO) algorithm and a Generalized Additive Model (GAM) using penalty splines (p-splines). The performance of these models was measured using statistical quality parameters and the results obtained indicate that the MAG provides more accurate estimates.