Cover Image

Machine learning classifiers for modeling soil characteristics by geophysics investigations: a comparative study

To design geotechnical structures efficiently, it is important to examine soil's physical properties. Therefore, classifying soil with respect to geophysical parameters is an advantageous and popular approach. Novel, quick, cost, and time effective machine learning techniques can facilitate thi...

Full description

Saved in:
Bibliographic Details
Main Authors: Lim, Chee Soon, Mohamad, Edy Tonnizam, Motahari, Mohammad Reza, Armaghani, Danial Jahed, Saad, Rosli
Format: Article
Language:English
Published: MDPI 2020
Subjects:
TA Engineering (General). Civil engineering (General)
Online Access:http://eprints.utm.my/id/eprint/91631/1/LimCheeSoon2020_MachineLearningClassifiersforModelingSoil.pdf
http://eprints.utm.my/id/eprint/91631/
http://dx.doi.org/10.3390/APP10175734
Tags: Add Tag
No Tags, Be the first to tag this record!
id my.utm.91631
record_format eprints
spelling my.utm.916312021-07-14T08:18:52Z http://eprints.utm.my/id/eprint/91631/ Machine learning classifiers for modeling soil characteristics by geophysics investigations: a comparative study Lim, Chee Soon Mohamad, Edy Tonnizam Motahari, Mohammad Reza Armaghani, Danial Jahed Saad, Rosli TA Engineering (General). Civil engineering (General) To design geotechnical structures efficiently, it is important to examine soil's physical properties. Therefore, classifying soil with respect to geophysical parameters is an advantageous and popular approach. Novel, quick, cost, and time effective machine learning techniques can facilitate this classification. This study employs three kinds of machine learning models, including the Decision Tree, Artificial Neural Networks, and Bayesian Networks. The Decision tree models included the chi-square automatic interaction detection (CHAID), classification and regression trees (CART), quick, unbiased, and efficient statistical tree (QUEST), and C5; the Artificial Neural Networks models included Multi-Layer Perceptron (MLP) and Radial Basis Function (RBF); and BN models included the Tree Augmented Naïve (TAN) and Markov Blanket, which were employed to predict the soil classifications using geophysics investigations and laboratory tests. The performance of each model was assessed through the accuracy, stability and gains. The results showed that while the BAYESIANMARKOV model achieved the highest overall accuracy (100%) in training phase, this model achieved the lowest accuracy (34.21%) in testing phases. Thus, this model had the worst stability. The QUEST had the second highest overall training accuracy (99.12%) and had the highest overall testing accuracy (94.74%). Thus, this model was somewhat stable and had an acceptable overall training and testing accuracy to predict the soil characteristics. The future studies can use the findings of this paper as a benchmark to classify the soil characteristics and select the best machine learning technique to perform this classification. MDPI 2020-09 Article PeerReviewed application/pdf en http://eprints.utm.my/id/eprint/91631/1/LimCheeSoon2020_MachineLearningClassifiersforModelingSoil.pdf Lim, Chee Soon and Mohamad, Edy Tonnizam and Motahari, Mohammad Reza and Armaghani, Danial Jahed and Saad, Rosli (2020) Machine learning classifiers for modeling soil characteristics by geophysics investigations: a comparative study. Applied Sciences (Switzerland), 10 (17). pp. 1-21. ISSN 2076-3417 http://dx.doi.org/10.3390/APP10175734 DOI:10.3390/APP10175734
institution Universiti Teknologi Malaysia
building UTM Library
collection Institutional Repository
continent Asia
country Malaysia
content_provider Universiti Teknologi Malaysia
content_source UTM Institutional Repository
url_provider http://eprints.utm.my/
language English
topic TA Engineering (General). Civil engineering (General)
spellingShingle TA Engineering (General). Civil engineering (General)
Lim, Chee Soon
Mohamad, Edy Tonnizam
Motahari, Mohammad Reza
Armaghani, Danial Jahed
Saad, Rosli
Machine learning classifiers for modeling soil characteristics by geophysics investigations: a comparative study
description To design geotechnical structures efficiently, it is important to examine soil's physical properties. Therefore, classifying soil with respect to geophysical parameters is an advantageous and popular approach. Novel, quick, cost, and time effective machine learning techniques can facilitate this classification. This study employs three kinds of machine learning models, including the Decision Tree, Artificial Neural Networks, and Bayesian Networks. The Decision tree models included the chi-square automatic interaction detection (CHAID), classification and regression trees (CART), quick, unbiased, and efficient statistical tree (QUEST), and C5; the Artificial Neural Networks models included Multi-Layer Perceptron (MLP) and Radial Basis Function (RBF); and BN models included the Tree Augmented Naïve (TAN) and Markov Blanket, which were employed to predict the soil classifications using geophysics investigations and laboratory tests. The performance of each model was assessed through the accuracy, stability and gains. The results showed that while the BAYESIANMARKOV model achieved the highest overall accuracy (100%) in training phase, this model achieved the lowest accuracy (34.21%) in testing phases. Thus, this model had the worst stability. The QUEST had the second highest overall training accuracy (99.12%) and had the highest overall testing accuracy (94.74%). Thus, this model was somewhat stable and had an acceptable overall training and testing accuracy to predict the soil characteristics. The future studies can use the findings of this paper as a benchmark to classify the soil characteristics and select the best machine learning technique to perform this classification.
format Article
author Lim, Chee Soon
Mohamad, Edy Tonnizam
Motahari, Mohammad Reza
Armaghani, Danial Jahed
Saad, Rosli
author_facet Lim, Chee Soon
Mohamad, Edy Tonnizam
Motahari, Mohammad Reza
Armaghani, Danial Jahed
Saad, Rosli
author_sort Lim, Chee Soon
title Machine learning classifiers for modeling soil characteristics by geophysics investigations: a comparative study
title_short Machine learning classifiers for modeling soil characteristics by geophysics investigations: a comparative study
title_full Machine learning classifiers for modeling soil characteristics by geophysics investigations: a comparative study
title_fullStr Machine learning classifiers for modeling soil characteristics by geophysics investigations: a comparative study
title_full_unstemmed Machine learning classifiers for modeling soil characteristics by geophysics investigations: a comparative study
title_sort machine learning classifiers for modeling soil characteristics by geophysics investigations: a comparative study
publisher MDPI
publishDate 2020
url http://eprints.utm.my/id/eprint/91631/1/LimCheeSoon2020_MachineLearningClassifiersforModelingSoil.pdf
http://eprints.utm.my/id/eprint/91631/
http://dx.doi.org/10.3390/APP10175734
_version_ 1706956973891649536
score 13.223943

Similar Items