Keynote - Dr Hans Georg Zimmermann at IntelliSys 2015 - Recurrent Neural Networks

Talk Title: System Identification and Forecasting with Recurrent Neural Networks Abstract: Observations and measured data belong to the past – decision support and optimal control are based on an understanding of the future. The intermediate link between past and future is system identification which has to show its correctness in its forecasting power. If we have no apriori knowledge about the linearity of our system of interest we should start the analysis with a class of universal approximation models.