Knowledge-Based Neural Networks for Modelling Time Series

Abstract

Various methods exist for extracting rules from data for classification purposes. We propose a new method for initializing a neural network used for time series modelling and prediction. We extract binary rules from a real valued time series and encode them into a neural network using an adaptation of KBANN. We test the method on the Lorenz system as well as on real world data in the form of a seismic time series. Results show that the method is successful in extracting and encoding prior knowledge. For the Lorenz system training time was halved and better generalization performance in the form of a lower mean squared error was obtained. Better training time and generalization performance was also obtained for the seismic time series.