 1. The paper focuses on predicting energy consumption in public buildings using neural networks. 2. The authors use data from an energy monitoring system at the University of Granada and compare the performance of non-linear autoregressive, AR, and non-linear autoregressive neural network with exogenous inputs, NRX. 3. Results show that both NR and NRX models are effective for predicting energy consumption, but exogenous data can improve accuracy. This article was authored by Luis Gonzaga Baca Ruiz, Manuel Pagalajar Cuellar, Miguel Delgado Calvafloris, and others.