New Model for Predicting Soil Resilient Modulus
Author Information
Author(s): Xiangfeng Duan
Primary Institution: Imperial College London, London, UK
Hypothesis
Can a new model improve the prediction accuracy of soil resilient modulus using machine learning techniques?
Conclusion
The BKA-XGBOOST model significantly outperforms traditional methods and other machine learning models in predicting soil resilient modulus.
Supporting Evidence
- The proposed model achieved a determination coefficient (R2) of 0.995.
- The mean absolute error (MAE) of the model was 0.975 MPa.
- BKA-XGBOOST was compared with nine other models and showed superior performance.
- The model was developed to be user-friendly for engineers.
Takeaway
This study created a new way to predict how strong soil is when it's under pressure, which helps in building better roads.
Methodology
The study used a hybrid model combining black-winged kite algorithm and extreme gradient boosting to analyze 2813 soil data points.
Limitations
The model's generalizability is limited due to the small variety of soil types used in the experiments.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website