This paper describes a strategy for building a predictive model for actual
complex time series. Time series data of temperature fluctuations observed
in a blast furnace for iron-making are taken as an example. Chaotic feature
s of the data are investigated with diagnostic algorithm for instability an
d parallelism of neighboring trajectories in phase space reconstructed from
the time series data. Stationarity of the data is examined with diagnostic
algorithm based on the KM2O-Langevin equations developed by Okabe. A short
time series for which no control actions were taken to the plant during me
asurement is diagnosed as possibly low-dimensional chaos, while for a long
time series including many control actions during measurement, determinism
is less visible and its predicted behavior exhibits a scaling property simi
lar to self-affine random noise. Characteristic exponents are estimated fro
m the scaling properties of the prediction error as a function of the predi
ction-time interval. Such information is exploited as prior knowledge for d
esigning a generalized Gaussian radial basis function network as a predicto
r The performance of the network is improved when linear algebraic polynomi
als are added to the network. The characteristic exponents estimated are us
ed as reliability indices of forecasting future trends of the data. (C)2000
Elsevier Science B.V. All rights reserved.