Contribution to Optimizing an Echo-State Based Cellular Neural Network

Forecasting is a very important factor in management and decision-making in many fields ranging from finance to engineering. In this book we present the use of neural network in the Time-series Forecasting process. The model we have chosen to base our study is the Celular Neural Network (CNN), to which we have associated the learning paradigm called Echo State Network (ESN). Both of these concepts have very good individual properties that we have combined and applied to the Forecasting of robust Time-series. We also discuss the optimization aspect by studying the impact of some parameters of the final model on the performance of the system with regard to the forecasting error.