In this study, a machine learning technique is applied to predict earthquake induced liquefaction. This method is a fairly new pattern recognition tool known as a support vector machine (SVM). SVM has been successfully applied in many applications, but it is less widely applied in the geotechnical field. This paper presents an SVM-based approach for predicting liquefaction. The model was trained and tested on a data set comprising 466 field records of liquefaction performance and CPT measurements. A grid search method with k-fold crossvalidation is also used to verify the feasibility. Compared with the Artificial Neural Network (ANN)–based method, the SVM-based method has the advantage of increased accuracy and simpler operation. The experimental results show that an SVM can increase the classification accuracy rate to a standard of 98.7%.
Soil liquefaction is a vital topic of concern in the field of geotechnical engineering. Experts and scholars in this field seek methods to correctly determine whether liquefaction will occur at a site. The methods developed over the past several decades for evaluating soil liquefaction primarily involve simple empirical methods using on-site test data. However, the great uncertainties inherent in earthquake mechanisms and soil properties complicate the selection of a suitable empirical formula for conducting regression analysis. Therefore, experts and scholars have attempted to discover an analytical method that is simpler, and better able to accurately determine soil liquefaction than traditional empirical formulae. In recent years, data processing and analytical ability have greatly increased and the cone penetration test (CPT) has the advantage of being a fast, continuous and accurate measurement of soil parameters. Meanwhile, related testing data has continued to accumulate. Therefore, the potential of applying CPT to liquefaction research has grown significantly (Roberston and Wride, 197; Roberston and Fear, 1996; Seed and Idriss, 1971).
