Track: Machine Learning
Adoption of a good estimation model for the prediction of subsoils properties before the commencement of a construction project, or at the preliminary stage of project planning is highly imperative. This will mitigate the most unexpected costs incurred during construction which are mostly geotechnical in nature This research aims to use Machine Learning(ML) tools such as Multiple Linear Regression(MLR)Artificial Neural Network (ANN), Support Vector Machine (SVM), Random Forest (RF) and M5 Tree (M5P) in geotechnical Engineering with a view to correlate Optimum Moisture Content(OMC),Maximum Dry Density (MDD) and Soaked California Bearing Ratio (SCBR) and Unsoaked California Bearing Ratio (USCBR) from the laboratory by conducting tests on 480 disturbed soil samples The principal component analysis (PCA) was used to examine Collinearity problem among the data set. The actual and predicted values of the machine learning (ML) models using root mean square error (RMSE) showed a varied values of RMSE and Coefficient of determination (R2). The results varied between 0.04 -0.92 and 0.02 – 0.93 for OMC and MDD while SCBR and UNSCBR ranges from 0.16 to 0.94 and 0.01 to 0.92 f respectively. From the fore going RF gave the least values of RMSE as 101.63 and 1.67 and the highest value of R2 as 0.92 and 0.93. The results showed that these models had R2 values greater than 90% and the variation of error between the observed and the predicted values of estimated geotechnical parameters was less than ±2. Its concluded that these models will be useful for preliminary design of Civil engineering infrastructure in Ekiti-State, South Western, Nigeria.
Keywords: Machine Learning, Geotechnical indices, Models, Coefficient of determination, Root Mean Square Error