Geotechnics for Sustainable Infrastructure Development pp 1287-1294 | Cite as
Development of Locally Specified Soil Stratification Method with CPT Data Based on Machine Learning Techniques
Cone Penetration Test (CPT) provides us with subsurface information with high resolution and good accuracy, which does not confirm the types of in-situ geomaterials directly. The engineering experience-driven classification charts or tables are usually used when CPT data is applied for soil stratification. However, these charts or tables have an inherent limitation that they were derived merely based on the given field experiences, which indicates that these cannot represent the engineering characteristic of all the soils in the world. This study proposes that the development of locally modified CPT-based soil classification methods can be performed with machine learning techniques. The results show that, using the simply trained algorithm with sufficient training data, the locally specified soil stratification is possible with high accuracy.
Keywordsstratification cone penetration test machine learning data mining decision tree model
Unable to display preview. Download preview PDF.
- Ian H.W. (2011). DATA MINING: Practical Machine Learning Tools and Techniques. Elsevier.Google Scholar
- Quinlan, J. R. (1986). Induction of decision trees. Machine Learning, 1(1), pp 81-106.Google Scholar
- Quinlan, J.R. (2008). Top 10 algorithms in data mining. Knowledge and Information Systems, 14(1), pp 1-37Google Scholar
- Robertson, P.K. (2016). Cone penetration test – based soil behaviour type classification system – an updated. Canadian Journal of Geotech, 53, pp 1910-1927.Google Scholar