The feed-forward neural network was used to simulate the behaviour of soil
samples in uniaxial strain conditions, i.e. to predict the oedometer test r
esults only on the basis of the basic soil properties. Artificial neural ne
twork was trained using the database of 217 samples of different cohesive s
oils from various locations in Slovenia. Good agreement between neural netw
ork predictions and laboratory test results was observed for the test sampl
es. This study confirms the link between basic soil properties and stress-s
train soil behaviour and demonstrates that artificial neural network succes
sfully predicts soil stiffness in uniaxial strain conditions. The compariso
n between the neural network prediction and empirical formulae shows that t
he neural network gives more accurate as well as more general solution of t
he problem. (C) 2001 Civil-Comp Ltd and Elsevier Science Ltd. All rights re
served.