Application of BP neural network in prediction of compression index of soil

In order to acquire deformation parameters based on many routine physical parameters, a prediction of compression index was carried out with BP neural network based on the testing data of soil in several engineering sites. Taking plasticity index, water content, void ration and density of soil as primary influence factors, the prediction model of compression index based on BP neural network was obtained. The results show that the relative error of fitting value of compression index compared with the observed value for 49 groups of independent variables training BP neural network model is from -3.513 938 0% to 1.570 422 5%, and the average value of absolute value of relative error is 0.915 48%. And the relative error of fitting value of compression index compared with the observed value for 10 groups of independent variables validating BP neural network model is from -1.805 521 0% to 6.012 417 3%, and the average value of absolute value of relative error is 3.329 40%. Therefore, the prediction model of compression index with BP neural network based on 4 routine physical parameters is doable.