Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Novel approach for soil classification using machine learning methods
Abstract In this study, we have proposed a new classification method for determining different soil classes based on three machine learning approaches, namely: support vector classification (SVC), multilayer perceptron (MLP), and random forest (RF) models. For the development of models, we have used a database of 4888 soil samples obtained from Vietnam projects. In the model’s study, 15 soil properties factors (variables) have been selected as input parameters for classifying soil samples into 5 soil classes: lean clay (CL), elastic silt (MH), fat clay (CH), clayey sand (SC), and silt (ML). To evaluate and analyze the results quantitatively and qualitatively, various methods such as learning curve (time and number of training samples), confusion matrix, and several statistical metrics such as precision, recall, accuracy, and F1-score were used. Results indicated that performance of all the three models (average accuracy score = 0.968) is good but of the SVC model (accuracy score = 0.984) is best in accurate classification of soils.
Novel approach for soil classification using machine learning methods
Abstract In this study, we have proposed a new classification method for determining different soil classes based on three machine learning approaches, namely: support vector classification (SVC), multilayer perceptron (MLP), and random forest (RF) models. For the development of models, we have used a database of 4888 soil samples obtained from Vietnam projects. In the model’s study, 15 soil properties factors (variables) have been selected as input parameters for classifying soil samples into 5 soil classes: lean clay (CL), elastic silt (MH), fat clay (CH), clayey sand (SC), and silt (ML). To evaluate and analyze the results quantitatively and qualitatively, various methods such as learning curve (time and number of training samples), confusion matrix, and several statistical metrics such as precision, recall, accuracy, and F1-score were used. Results indicated that performance of all the three models (average accuracy score = 0.968) is good but of the SVC model (accuracy score = 0.984) is best in accurate classification of soils.
Novel approach for soil classification using machine learning methods
Nguyen, Manh Duc (Autor:in) / Costache, Romulus (Autor:in) / Sy, An Ho (Autor:in) / Ahmadzadeh, Hassan (Autor:in) / Van Le, Hiep (Autor:in) / Prakash, Indra (Autor:in) / Pham, Binh Thai (Autor:in)
2022
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
BKL:
56.00$jBauwesen: Allgemeines
/
38.58
Geomechanik
/
38.58$jGeomechanik
/
56.20
Ingenieurgeologie, Bodenmechanik
/
56.00
Bauwesen: Allgemeines
/
56.20$jIngenieurgeologie$jBodenmechanik
RVK:
ELIB18
Machine learning in soil classification
| Tema Archiv | 2005
CPT-Based Soil Classification through Machine Learning Techniques
| British Library Conference Proceedings | 2023