The group method of data handling (GMDH) algorithm presented by A. C. Ivakh
nenko and colleagues is an heuristic self-organization method. It establish
es the input-output relationship of a complex system using a multilayered p
erception-type structure that is similar to a feed-forward multilayer neura
l network. This study provides a step towards understanding and evaluating
a role for GMDH in the investigation of the complex rainfall-runoff process
es in a heterogeneous watershed in Taiwan. Two versions of the revised GMDH
model are implemented: a stepwise regression procedure and a recursive for
mula. Eleven typhoon events in the Shen-cei Creek watershed, Taiwan, are us
ed to build the model and verify its usefulness. The prediction results of
the revised GMDH models and the instantaneous unit hydrograph (IUH) model a
re compared. Based on the criteria of forecasting precision and the rate an
d time of peak error, a much better performance is obtained with the revise
d GMDH models. Copyright (C) 1999 John Wiley & Sons, Ltd.