Title :
Interpretation of recurrent neural networks
Author :
Pedersen, Morten With ; Larsen, Jan
Author_Institution :
Dept. of Math. Modelling, Tech. Univ. Lyngby, Denmark
Abstract :
This paper addresses techniques for interpretation and characterization of trained recurrent nets for time series problems. In particular, we focus on assessment of effective memory and suggest an operational definition of memory. Further we discuss the evaluation of learning curves. Various numerical experiments on time series prediction problems are used to illustrate the potential of the suggested methods
Keywords :
learning (artificial intelligence); recurrent neural nets; time series; effective memory; learning curves; recurrent neural networks; time series prediction problems; trained recurrent nets; Control systems; Cost function; Feedforward systems; Fractals; Mathematical model; Neurofeedback; Phase measurement; Recurrent neural networks; Stability analysis; System identification;
Conference_Titel :
Neural Networks for Signal Processing [1997] VII. Proceedings of the 1997 IEEE Workshop
Conference_Location :
Amelia Island, FL
Print_ISBN :
0-7803-4256-9
DOI :
10.1109/NNSP.1997.622386
