Civil and Geoengineering Letters Editorial Office info@jcivilgeo.com mEDRA 20260525 DOI Registration 01

06 10.22034/CGEL.3.1.e100052 https://jcivilgeo.com/article/e100052 Civil and Geoengineering Letters mEDRA 01 01 3115-8129 Civil and Geoengineering Letters Editorial Office IR 01 3115-8129 JB 3 1 01 202605 1 1 A01 Alireza Akbari Akbari, Alireza Alireza Akbari 2 A01 Ali Fatemi Fatemi, Ali Ali Fatemi 01 eng 202605 03 00 This study evaluates the effectiveness of three machine learning models such as Artificial Neural Network (ANN), Support Vector Machine (SVM), and Logistic Regression (LR), for the classification of fine-grained soils based on index properties. A dataset comprising 1,257 samples was compiled through an extensive literature review. Each sample included key index parameters such as liquid limit (LL), plastic limit (PL), plasticity index (PI), and percentage of fines. The methodology involved preprocessing the data, normalizing features, and dividing the dataset into training (70%) and testing (30%) sets. Models were trained and validated using k-fold cross-validation (k=10), and their performance was assessed using accuracy, precision, recall, F1-score, and confusion matrices. The ANN model achieved the highest classification accuracy of 91.6%, followed by SVM at 88.3%, and LR at 84.9%. The mean squared error (MSE) for ANN, SVM, and LR were calculated as 0.078, 0.102, and 0.139, respectively. The ANN model also achieved the highest F1-score (0.92), indicating superior consistency in classification. These results highlight the potential of machine learning models, particularly ANN, for the reliable and automated classification of fine-grained soils using basic index properties. 20 Fine-grained soils; Artificial neural network; Support vector machine; Logistic regression; Soil index properties 01 00 20260218 02 00 20260312 2026 Ali Fatemi