Abstract: The lateral load capacity of a single pile is an important parameter in the civil engineering design practice. To establish the nonlinear relationships among potential influencing factors, a Support Vector Machine (SVM) model for predicting the capacity of laterally loaded piles was established. By learning from a small number of samples, this model can accurately predict the prediction samples that only know the potential influencing factors. The accuracy of the model was verified by comparing the prediction results of the SVM model and BP neural network model. Both the mean relative error and the root mean square error are used to evaluate the overall performance and stability of each model. The high reliability of the predicted results was verified by the confidence interval, and all the predicted values are within the confidence interval of 90%, 95% and 97%. The Maximum Relative Error (MXAE) of SVM is 5.41%, the MRE is 2.81%, and the RMSE is 2.0278. Compared with the prediction results of BP neural network, the SVM results are more accurate. The SVM model is therefore feasible in predicting the bearing capacity of laterally loaded piles and provides a new method for its acquisition.

Key words: support vector machines, laterally loaded, pile bearing capacity, prediction model