Neural networks for large financial crashes forecast

The aim of this work is to examine how neural networks can be used for solving the problem of the forecast of large financial crashes due to the presence of speculative bubbles. Some microeconomic theories have been developed for the explanation of a bubble due to a cooperation among the investors. This behaviour can be detected by the presence of self-similarity in the indexes series near the crash time leading to a differential equation and thus to a dynamical system description, well suitable by a neural network approach.