Non-linear system identification of an autocatalytic reactor using least squares support vector machines
The concepts behind support vector machines are very interesting both in theory and in practice, as they are based on a universal constructive learning procedure derived from the statistical learning theory developed by Vapnik. In this paper, their application to time series modelling is considered by means of simulated data from an autocatalytic reactor. In particular, reconstruction methods from non-linear dynamics are used to define a state space model for the process. Multivariate embedding techniques are compared to scalar embedding with respect to modelling.