Application of extreme gradient boosting for predicting standard penetration test N-values from cone penetration test data

The article delves into the critical role of the Standard Penetration Test (SPT) and Cone Penetration Test (CPT) in geotechnical engineering, emphasizing their importance in characterizing subsurface stratigraphy and evaluating geotechnical properties. It highlights the historical and contemporary methods used to correlate SPT N-values with CPT data, noting the limitations of traditional approaches. The study introduces a novel CPT-SPT transformation model developed using Extreme Gradient Boosting (XGBoost), which demonstrates superior predictive capability compared to conventional models and other machine learning techniques such as Random Forests, Back-Propagation Artificial Neural Networks, and Support Vector Machines. The article provides a detailed analysis of the model's performance, feature importance, and interpretability, using Shapley Additive Explanations (SHAP) to elucidate the contribution of each input variable. The results indicate that the XGBoost model achieves the highest coefficient of determination (R²) and the lowest root mean square error (RMSE), making it a robust tool for predicting SPT N-values from CPT data. The article also discusses the implications of these findings for geotechnical engineering practices, suggesting that the XGBoost model can enhance the accuracy and reliability of soil characterization and design processes. Furthermore, the study establishes a high-quality side-by-side CPT-SPT database, ensuring minimal horizontal variability and providing a reliable foundation for future research and applications.