Assessing the Performance of Machine Learning Algorithms for Soil Classification Using Cone Penetration Test Data

Conventional soil classification methods are expensive and demand extensive field laboratory work. This research evaluates the efficiency of various machine learning (ML) algorithms in classifying soils based on Robertson’s behavioral types. study employs 4 ML algorithms, including artificial neural network (ANN), random forest (RF), support vector (SVM), decision trees (DT), to classify from 232 cone penetration test (CPT) datasets. The datasets were randomly split into training testing train models. Metrics such as overall accuracy, sensitivity, precision, F1_score, confusion matrices provided quantitative evaluations each model. Our analysis showed that all models accurately classified most soils. SVM model achieved highest accuracy 99.84%, while ANN an 98.82%. RF DT scores 99.23% 95.67%, respectively. Additionally, evaluation metrics indicated high scores, demonstrating performed well. exhibited outstanding performance both majority minority classes, lower sensitivity F1_score for class. Based these results, we conclude can be integrated software programs rapid accurate classification.