Identificación de Suscriptores en Riesgo de Abandono Utilizando Modelamiento de Aprendizaje de Máquina para Predicción por Clasificación para una Empresa de Medicina Prepagada del Ecuador.

Abstract

The present research addresses the issue of identifying subscribers at risk of churn in a prepaid medical company based in Ecuador. The main objective of this study is to develop and implement a classification prediction system based on machine learning techniques, allowing the identification of subscribers with a high probability of abandoning the services offered by the company. In the literature review phase, previous research related to customer retention in the health context and the use of machine learning algorithms for churn prediction in different industries is analyzed. Key theoretical concepts such as classification algorithms, modeling techniques, and peculiarities of the prepaid medical system in Ecuador are also explored. The proposed methodology involves the collection and analysis of historical subscriber data, including demographic information, patterns of medical service usage, and prior churn behavior. Various machine learning techniques, such as logistic regression, support vector machines, and decision trees, will be employed to build and compare predictive models. Model evaluation will be conducted using performance metrics such as accuracy, sensitivity, and specificity.