Assessment of Feature Selection for Liquefaction Prediction Based on Recursive Feature Elimination

This paper presents a machine learning model using random forest (RF) algorithm with the recursive feature elimination (RFE) for soil liquefaction prediction. The prediction is tested on 253 CPT-based field data from different earthquakes. RFE, which one of selection methods, was adopted eliminating irrelevant features in mentioned dataset, and then performance RFE-RF (i.e., determined by RFE method) RF models all variables were compared terms their matrices. primary focus this study to investigate effectiveness approach, therefore raw that benchmark dataset used compare RFE-RF. result showed approach improved overall accuracy