|
Neural Network vs. Autoregressive Classical Mathematical Models for Short-Term Load Forecasting Resumen: This paper presents an analysis of one of the forecasting models running at the National Transport Operator System, which is based on both autoregressive mathematical models and neural network techniques of artificial intelligence. The results of this research show which models are better, depending on the available data and which periods are more accurately forecasted by each model and provide valid criteria to choose one or the other. Firstly, to be able to know which method would be better suited to a particular electric system or data set it has been analyzed many techniques such as: statistical methods, complex artificial intelligence systems, and many others, to be implemented in the application for the electric system comparing its performance and accuracy for a definite period. Secondly, different criteria and data filtering techniques have been used to obtain the optimal results. |