A Prediction Method of Non-Stationary Time Series Data by Using a Modular Structured Neural Network

Eiji WATANABE  Noboru NAKASAKO  Yasuo MITANI  

Publication
IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences   Vol.E80-A   No.6   pp.971-976
Publication Date: 1997/06/25
Online ISSN: 
DOI: 
Print ISSN: 0916-8508
Type of Manuscript: Special Section PAPER (Special Section on Signal Processing Theories and Applications Based on Modelling of Nonstationary Processes)
Category: 
Keyword: 
AR model with time varying paramenters,  time series prediction,  modular structured neural networks,  additive learning ability,  

Full Text: PDF(480.6KB)>>
Buy this Article




Summary: 
This paper proposes a prediction method for non-stationary time series data with time varying parameters. A modular structured type neural network is newly introduced for the purpose of grasping the changing property of time varying parameters. This modular structured neural network is constructed by the hierarchical combination of each neural network (NNT: Neural Network for Prediction of Time Series Data) and a neural network (NNW: Neural Network for Prediction of Weights). Next, we propose a reasonable method for determination of the length of the local stationary section by using the additive learning ability of neural networks. Finally, the validity and effectiveness of the proposed method are confirmed through simulation and actual experiments.