Artificial neural networks and anticipatory systems

Neural networks are structures consisting of highly interconnected elementary computational units. The network map input patterns into output patterns. These devices are called neural not because they model the nervous systems but because they are inspired by them. Due to their features (adaptive behaviour, learning process, robustness,...) , artificial neural networks have interested the anticipatory systems field of research. In particular, we will examine their interests in the field of time series prediction and system identification.