Modelling of Seismic Liquefaction Using Classification Techniques

Liquefaction susceptibility of soil is a complex problem due to non-linear behaviour of soil and its physical attributes. The assessment of liquefaction potential is commonly assessed by the in-situ testing methods. The classification problem of liquefaction is non-linear in nature and difficult to model considering all independent variables (seismic and soil properties) using traditional techniques. In this study, four different classification techniques, namely Fast k-NN (F-kNN), Naïve Bayes Classifier (NBC), Decision Forest Classifier (DFC), and Group Method of Data Handling (GMDH), were used. The SPT-based case record was used to train and validate the models. The performance of these models was assessed using different indexes, namely sensitivity, specificity, type-I error, type-II error, and accuracy rate. Additionally, receiver operating characteristic (ROC) curve were plotted for comparative study. The results show that the F-kNN models perform far better than other models and can be used as a reliable technique for analysis of liquefaction susceptibility of soil.