[en] Artificial neural networks seem to be a promising tool to perform classification. To do it, the network is trained in a proper way, i.e. both normal and abnormal signals are shown to the network. Using a large enough sample set, after the training period the network is able to distinguish normal and abnormal signals. When a new signal is shown, a network assigns it to either normal or abnormal class, using the information inherently extracted from the sample set. The method is under development but early and promising results indicate the applicability of the tool in real situations. (author)