Correlation analysis between the collapsibility of loess and physicomechanical indices in Longdong region and prediction of the collapsibility coefficient based on machine learning

) of 0.910 and 0.900 for the training set and testing set, respectively. The prediction accuracy for identifying whether loess is collapsible or nocollapsible reaches 92.7%. SHAP analysis reveals that the influence trends of various factors on model predictions are consistent with theoretical cognition, and soils with a depth less than 11 m and a void ratio greater than 1.02 have a high collapsibility tendency. The results and conclusions of this study can provide a theoretical basis for subsequent engineering construction in the Longdong area.