Application of Random Forest Regression in the Prediction of Ultimate Bearing Capacity of Strip Footing Resting on Dense Sand Overlying Loose Sand Deposit

Abstract

The paper presents the prediction of the ultimate bearing capacity of the strip footing resting on layered soil (dense sand overlying loose sand) using random forest regression (RFR). In this study, 181 data collected from literature were used. 71 % of the total data was randomly selected for training the model and the rest of the data were utilized for the testing purpose. The various input parameters were friction angle of the dense sand layer (f1), friction angle of the loose sand layer (f2), unit weight of the dense sand layer (g1), unit weight of the loose sand layer (g2), ratio of the thickness of the dense sand layer below base of the footing to the width of footing (H/B), ratio of the depth of the footing to the width of the footing (D/B) and (H+D)/B. Ultimate bearing capacity was the output in this study. Performance measures were used in order to make the comparison with the artificial neural network (ANN) and M5P model tree. The result of this study revealed that the performance of the RFR was superior to M5P and ANN. The results of the sensitivity analysis reveals that the unit weight and the friction angle of the loose sand layer were the most important parameters affecting the output ultimate bearing capacity of the strip footing resting on the layered soils.

Keywords

• random forest regression
• ultimate bearing capacity
• layered sand
• m5p model tree
• artificial neural network
• sensitivity analysis