Civil Engineering and Architecture Vol. 11(4), pp. 2012 - 2032
DOI: 10.13189/cea.2023.110425
Reprint (PDF) (2243Kb)

The Role of Hybrid Machine Learning for Predicting Strength Behavior of Sustainable Concrete

Bader Aldeen Ayman Almahameed ^1,*, Habibur Rahman Sobuz ^2
1 Maan University College, Al-Balqa Applied University, Jordan
² Department of Building Engineering and Construction Management, Khulna University of Engineering and Technology, Khulna-9203, Bangladesh

ABSTRACT

Researchers are actively seeking accurate models for predicting forecasting mechanical strength in response to the proliferation of novel mixtures of concrete and applications. Both linear and nonlinear regression, two types of empirical and statistical models, have seen extensive use. Sustainable concrete is made by introducing supplemental cement elements into concrete mixing, and it finds widespread use in sound attenuation, roofing, thermal insulation, varied tunneling, and geotechnical engineering. The effectiveness of this technology depends on its capacity to provide consistent products with predictable outcomes. In this article, we train and test our ML approaches and modeling using an experimental database comprised of relevant data obtained from numerous prior investigations. Through a new combination of the random forests (RF) model and the Bagging algorithm, this work introduces a hybrid ML model (RF-B) for forecasting the compressive strength of concrete. Bagging is an ensemble approach that aggregates the predictions of numerous models that were each fit to a separate subset of a training dataset. As a second example, Support Vector Regression (SVR) was created to help in finding the activities of parameters in connection to one another in order to forecast the robustness of machine learning models. Multivariate analysis is also another way of reading the data accumulated with a determination coefficient of roughly 0.6. The decision tree regression showed two iterations and R² values are 0.7453 and 0.7737 respectively. The cement percentage, density for oven dry conditions, w/c ratio, and additive usage are all used as input factors in the predictive models. Machine learning has many potential benefits for the construction industry, including cost savings, time savings, and less labor intensity. The statistical and graphical representation of contributors and countries in this study can facilitate the development of collaborative projects and the trading of novel ideas and approaches among scholars.

KEYWORDS
Sustainable Concrete, Random Forests, Bagging Algorithm, Machine Learning, Support Vector Regression, Multivariate Analysis

Cite This Paper in IEEE or APA Citation Styles
(a). IEEE Format:
[1] Bader Aldeen Ayman Almahameed , Habibur Rahman Sobuz , "The Role of Hybrid Machine Learning for Predicting Strength Behavior of Sustainable Concrete," Civil Engineering and Architecture, Vol. 11, No. 4, pp. 2012 - 2032, 2023. DOI: 10.13189/cea.2023.110425.

(b). APA Format:
Bader Aldeen Ayman Almahameed , Habibur Rahman Sobuz (2023). The Role of Hybrid Machine Learning for Predicting Strength Behavior of Sustainable Concrete. Civil Engineering and Architecture, 11(4), 2012 - 2032. DOI: 10.13189/cea.2023.110425.