Aplicación de modelos de aprendizaje automático y aprendizaje profundo, para la clasificación y predicción de servicios de la mesa de servicios OTRS de la Agencia de Regulación y Control de Electricidad

Abstract

Abstract: This study developed and implemented Machine Learning and Deep Learning models for the automatic classification and demand prediction of tickets in the OTRS service desk of the Agency for Regulation and Control of Electricity (ARCONEL). Based on the analysis of historical ticket data, the Knowledge Discovery in Databases (KDD) methodology was applied, enabling data selection, cleaning, transformation, and modeling processes aimed at identifying relevant patterns and anticipating periods of high demand. Classification and prediction models were trained and evaluated, and their results were integrated into a dashboard developed in Power BI, facilitating data visualization and validation to support decision-making. Additionally, a usability evaluation was conducted using the System Usability Scale (SUS) method, obtaining favorable results in terms of user acceptance and ease of use. The findings demonstrated that the application of artificial intelligence improved the efficiency of ticket management by optimizing response times and supporting operational planning within the Information Technology area.