Improving prediction of soil liquefaction using hybrid optimization algorithms and a fuzzy support vector machine

The phenomenon of soil liquefaction is one of the most complex and interesting fields in geotechnical earthquakes that has drawn the attention of many researchers in recent years. The present study used hybrid particle swarm optimization and genetic algorithms with a fuzzy support vector machine (FSVM) as the classifier for the soil liquefaction prediction problem. Fuzzy logic is used to decrease the outlier sensitivity of the system by inferring the importance of each sample in the training phase to increase the ability of the classifier's generalization. Using the appropriate combination of optimization algorithms, we can find the best parameters for the classifier during the training phase without the need for trial and error by the user due to the high accuracy of the classifier. The proposed approach was tested on 109 CPT-based field data from five major earthquakes between 1964 and 1983 recorded in Japan, China, the USA and Romania. Good results have been demonstrated for the proposed algorithm. Classification accuracy is 100% for the combination of the optimization algorithms with the FSVM classifier. The results show that the best kernel used in the training of the FSVM classifier is a radial basis function (RBF). According to the experimental results, the proposed algorithm can improve classification accuracy and that it is a feasible method for predicting soil liquefaction using the optimal parameters of the classifier with no user interface.