Faults Diagnostic using Hopfield Artificial Neural Network in front of Incomplete Data
Abstract
In this work, Hopfield Artificial Neural Network’s performance in faults diagnostic in industrial process is evaluated when there is missing data. The diagnostic of two classes with different levels of overlapping data is done. As main result, Hopfield has a good performance in the implemented tests getting over architectures like the Probabilistic Neural Network, that’s why it is a good option to use it in faults diagnostic.