The Prediction of Mechanical Properties of Cement Soil Based on PSO-SVM

The analysis of mechanical properties of cement soil is the problem concerned greatly by the design-constructors. This paper uses SVM to establish the nonlinear relation between the parameters and unconfined compressive strength of cement soil. At the same time, this paper uses PSO to get the global optimization of SVM, which avoids the blindness of artificial selection and improves the prediction accuracy of model. We apply PSO-SVM to the prediction of unconfined compressive strength, and then make a comparison with traditional BP-NN. The result shows that the prediction accuracy of PSO-SVM is much higher than BP-NN. Therefore, it is feasible and efficiency using PSO-SVM to predict the unconfined compressive strength of cement soil.