Predicción Del Consumo Energético En La Uleam Mediante Algoritmos de Aprendizaje Automático: Caso de Estudio

Due to the constant growth of inhabitants worldwide, urbanization and industrialization cause energy consumption to increase at an accelerated rate, generating importance in developing analysis of electrical energy consumption in order to better manage energy resources. Ecuador's energy panorama in 2024 is marked by a supply crisis that requires immediate actions in all economic and social sectors. Universities, being large consumers of electrical energy, face the challenge of balancing their operation with the commitment to sustainability. This research presents a case study at the Universidad Laica Eloy Alfaro de Manabí (ULEAM), where machine learning is used to predict energy consumption by applying data sciences. The data science methodology used in this study is OSEMN, in which 5 stages are applied. This research also focuses on identifying daily, weekly and seasonal patterns in energy use. Historical energy consumption data is analyzed, along with meteorological data, as well as the occupancy of individuals in the building or faculty. The performance of several algorithms is evaluated to select the most effective one. The results obtained in this research demonstrate that the Extreme Gradient Boosting (xgboost) and Random Forest algorithms generate a good level of performance for the data training process and for the testing process, which is used for energy consumption prediction.