Análisis de los factores que inciden en la selección del consumo en las estaciones de servicio de combustible en la ciudad de Machala a través de un modelo probabilístico de redes neuronales.

Abstract

This research analyzes the factors influencing fuel consumption choices at service stations in the city of Machala, from a microeconomic and applied perspective. Within a context characterized by regulated prices and subsidies, where price is not the sole competitive mechanism, the study examines how socioeconomic, operational, and perceptual variables influence the fueling decision. The methodology adopted was quantitative, based on a structured survey administered to 783 consumers. The data were subjected to descriptive and bivariate analysis and subsequently modeled using a multiclass classification Artificial Neural Network implemented in Python. This allowed for the identification of differentiated patterns according to consumption frequency, demonstrating that factors such as income level, price perception, waiting time, accessibility, and service quality significantly influence the consumer's decision. The results confirm that consumption behavior is heterogeneous and non-linear, validating the hypothesis that socioeconomic factors determine the selection of the fueling point and that these patterns can be modeled. predictively, providing local empirical evidence, contributing to the economic analysis of the fuel market and to strategic decision-making in the sector.