Appliaction of artificial neural networks to ground settlements

초록

Due to various reasons, theoretical prediction of ground settlement based on experimental data from laboratory misleads time-settlements behavior in the fields. In order to apply the artificial neural network to ground settlements, data based on the theoretical calculations were used beforehand for neural network training and testing. Thereafter the training and learning algorithms were further applied to the prediction of ground settlement of a field. With supplement of field data, the accuracy of prediction could be improved.