Understanding the geotechnical and geomechanical characteristics of erodible soils: a study incorporating soft computational modeling techniques

Johnbosco C. Egbueri () and Mohd Yawar Ali Khan
Additional contact information
Johnbosco C. Egbueri: Chukwuemeka Odumegwu Ojukwu University
Mohd Yawar Ali Khan: King Abdulaziz University

Environment, Development and Sustainability: A Multidisciplinary Approach to the Theory and Practice of Sustainable Development, 2024, vol. 26, issue 2, No 64, 4435-4466

Abstract: Abstract Soft computational algorithms enhance the understanding, prediction, and classification of engineering and environmental problems. In Nigeria, works that have used these techniques in modeling soil erodibility are scarce. In this paper, several soft computing methods were integrated to assess and model the geotechnical and geomechanical characteristics of gully soils in Southeast Nigeria. Standard methods were employed for soil geotechnical analysis. Relationships between geotechnical parameters were estimated using R-type hierarchical clustering (HC), principal component analysis (PCA), and factor analysis (FA). Soil erodibility in the area was classified using Q-type HC and K-means clustering (KMC) algorithms. Moreover, gradient descent-optimized multilayer perceptron (GD-MLP) and multiple linear regression (MLR) models were developed to simulate and predict the soil properties, including fines %, sand %, gravel %, plasticity index, cohesion, and friction angle. This research indicated that: (1) all the analyzed soils have moderate–high erodibility characteristics, with high erodibility tendency predominant; (2) the PCA, FA, and R-type HC effectively captured the relationships of the geotechnical variables; (3) the Q-type HC and KMC models produced moderate and high erodibility clusters at 1:4 ratios; and (4) with low modeling errors, the MLR and GD-MLP accurately predicted the soil properties. However, overall, the MLR models (with R2 range of 0.976–1.000) outperformed the GD-MLP models (with R2 range of 0.924–0.998 and area under curve range of 0.900–0.948). Although high model performances were recorded in this work, future studies are encouraged to advance its findings.

Keywords: Artificial neural network; Factor analysis; Geotechnical properties; K-means clustering; Multiple linear regression; Soil erosion (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
http://link.springer.com/10.1007/s10668-022-02890-7 Abstract (text/html)
Access to the full text of the articles in this series is restricted.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:endesu:v:26:y:2024:i:2:d:10.1007_s10668-022-02890-7

Ordering information: This journal article can be ordered from
http://www.springer.com/economics/journal/10668

DOI: 10.1007/s10668-022-02890-7

Access Statistics for this article

Environment, Development and Sustainability: A Multidisciplinary Approach to the Theory and Practice of Sustainable Development is currently edited by Luc Hens

More articles in Environment, Development and Sustainability: A Multidisciplinary Approach to the Theory and Practice of Sustainable Development from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().