A number of neural network models and training procedures for time series p
rediction have been proposed in the technical literature. These models stud
ied for different time-variant data sets have typically used uni-directiona
l computation flow or its modifications. In this study, on the contrary, th
e concept of bi-directional computational style is proposed and applied to
prediction tasks. A bi-directional neural network model consists of two sub
networks performing two types of signal transformations bi-directionally. T
he networks also receive complementary signals from each other through mutu
al connections. The model not only deals with the conventional future predi
ction task, but also with the past prediction, an additional task from the
viewpoint of the conventional approach. An improvement of the performance i
s achieved through making use of the future-past information integration. S
ince the coupling effects help the proposed model improve its performance,
it is found that the prediction score is better than with the traditional u
ni-directional method. The bi-directional predicting architecture has been
found to perform better than the conventional one when tested with standard
benchmark sunspots data. (C) 2001 Elsevier Science Ltd. All rights reserve
d.