Al-Shamiri, A. K., Kim, J. H., Yuan, T. F., & Yoon, Y. S. (2019).

Modeling the compressive strength of high-strength concrete: An

extreme learning approach. Construction and Building Materials, 208,
204-219.
In a study by Al-Shamiri et al. (2019), the authors explore the use of the Extreme Learning Machine
(ELM) algorithm to predict the compressive strength of high-strength concrete (HSC), a crucial parameter
in concrete design. Utilizing 324 experimental data records with inputs such as water, cement, fine
aggregate, coarse aggregate, and superplasticizer, the study compares ELM's performance against a back
propagation (BP) trained artificial neural network (ANN). The results reveal that the ELM model
significantly outperforms the BP-ANN in terms of prediction accuracy and computational efficiency,
evidenced by lower RMSE, MAE, and MAPE values, as well as higher correlation coefficient (R) and
Nash-Sutcliffe efficiency (NSE) scores. This demonstrates ELM's superior generalization capability and its
potential as a robust alternative to traditional methods in predicting HSC compressive strength.

Ahmad, W., Ahmad, A., Ostrowski, K. A., Aslam, F., Joyklad, P., &
Zajdel, P. (2021). Application of advanced machine learning approaches
to predict the compressive strength of concrete containing
supplementary cementitious materials. Materials, 14(19), 5762.
In the paper by Ahmad et al. (2021), the authors investigate the application of advanced machine learning
(ML) techniques to predict the compressive strength (CS) of concrete containing supplementary
cementitious materials (SCMs) such as fly ash and blast furnace slag. Previous research has established that
ML techniques can significantly enhance the accuracy and efficiency of predicting concrete properties, as
opposed to traditional empirical methods which are often time-consuming and less reliable. The study
employs a variety of supervised machine learning algorithms, including Bagging, AdaBoost, Gene
Expression Programming (GEP), and Decision Tree (DT), to develop predictive models for CS. These
models were trained and validated using a substantial dataset derived from existing literature,
encompassing 1030 data points and considering eight input variables: cement, FA, BFS, water,
superplasticizer, coarse aggregate, fine aggregate, and age. The effectiveness of each model was evaluated
using performance metrics such as the coefficient of determination (R²), mean absolute error (MAE), mean
square error (MSE), and root mean square error (RMSE). Among the techniques, the Bagging model
demonstrated superior predictive accuracy, achieving an R² value of 0.92. This study underscores the
potential of ML approaches in civil engineering, particularly for the intelligent design of concrete mixes
incorporating SCMs, thereby optimizing resource use and reducing experimental effort.

Song, H., Ahmad, A., Farooq, F., Ostrowski, K. A., Maślak, M.,
Czarnecki, S., & Aslam, F. (2021). Predicting the compressive strength
of concrete with fly ash admixture using machine learning algorithms.
Construction and Building Materials, 308, 125021.
The literature review outlines the increasing trend of utilizing machine learning (ML) algorithms in civil
engineering, particularly for predicting various mechanical properties of concrete. Highlighting the
versatility of ML, the review discusses the application of a range of algorithms such as Artificial Neural
 Networks (ANN), Decision Trees (DT), Support Vector Machines (SVM), Random Forest (RF), Gene
Expression Programming (GEP), and Deep Learning (DL) in prior studies. Examples cited include
predictions of shear strength, compressive strength, and interfacial bond strength of concrete using different
ML models. The significance of the present research lies in its focus on fly ash-based concrete (FAC) and
the development of a predictive model using ML algorithms, aiming to understand input parameter roles
and achieve accurate predictions of compressive strength. The review sets the context for the current
study's experimental work and methodology, providing a foundation for comparing the performance of
various ML approaches in predicting concrete properties.

Cao, R., Fang, Z., Jin, M., & Shang, Y. (2022). Application of machine
learning approaches to predict the strength property of geopolymer
concrete. Materials, 15(7), 2400.
The study conducted by Cao, R., Fang, Z., Jin, M., & Shang, Y. (2022) focuses on utilizing
machine learning (ML) approaches to enhance the prediction accuracy of compressive strength
(C-S) in geopolymer concrete (GPC), particularly based on fly ash (FA). Geopolymer concrete,
as an environmentally friendly alternative to Portland cement mixtures, has garnered attention,
but the accuracy of strength prediction requires improvement. The research employs support
vector machine (SVM), multilayer perceptron (MLP), and XGBoost (XGB) techniques to predict
C-S for GPC, comparing experimental and predicted results. The XGB model demonstrates
superior accuracy with an R2 value of 0.98 compared to SVM (0.91) and MLP (0.88). Through
statistical checks and k-fold cross-validation, the high precision of the XGB model is confirmed,
with lower errors (MAE, MSE, RMSE) indicating its efficacy. Sensitivity analysis reveals FA as
the most influential parameter on C-S prediction. The study underscores the potential of ML
techniques to streamline experimental work in predicting material properties, offering both
accuracy and efficiency in concrete research.

Moein, M. M., Saradar, A., Rahmati, K., Mousavinejad, S. H. G.,

Bristow, J., Aramali, V., & Karakouzian, M. (2023). Predictive models
for concrete properties using machine learning and deep learning
approaches: A review. Journal of Building Engineering, 63, 105444.
The paper by Moein et al. (2023) provides a comprehensive review of the application of machine
learning (ML) and deep learning (DL) techniques in predicting concrete mechanical properties.
Beginning with an introduction to artificial intelligence (AI), the review traces the evolution of
AI and its impact on various technological advancements. It then delves into the fundamentals of
ML, discussing key concepts such as supervised and unsupervised learning, along with popular
algorithms like support vector machines (SVM) and decision trees. The review also explores the
principles of DL, highlighting its advantages in handling unstructured data and extracting
complex features directly from raw data. Furthermore, it discusses the application of ML and DL
techniques in predicting concrete properties such as compressive strength, elastic modulus, and
shear strength, summarizing studies that evaluate the performance of different algorithms in this
domain. Finally, the paper offers a comparative analysis of these techniques and outlines future
research directions to address current knowledge gaps and challenges in the field, aiming to
assist construction material engineers and researchers in selecting suitable techniques for their
applications.

Ahmad, A., Ahmad, W., Chaiyasarn, K., Ostrowski, K. A., Aslam, F.,
Zajdel, P., & Joyklad, P. (2021). Prediction of geopolymer concrete
compressive strength using novel machine learning algorithms.
Polymers, 13(19), 3389.
The paper by Ahmad et al. (2021) investigates the prediction of geopolymer concrete
compressive strength (CS) using novel machine learning (ML) algorithms. The
methodology involves utilizing supervised ML algorithms and obtaining input parameters
from published literature sources. The study employs individual ML techniques such as
Artificial Neural Network (ANN), as well as ensemble methods like AdaBoost and
boosting, based on nine input parameters including Na2SiO3, NaOH, SiO2, Na2O,
NaOH molarity, and curing age, to predict CS. Python coding within the Anaconda
software environment facilitates model execution, with evaluation based on coefficient
correlation (R2) values. The paper elaborates on the workings of ANN, resembling
biological neural networks, and discusses boosting and AdaBoost algorithms,
highlighting their role in ensemble learning for improving prediction accuracy. Results
demonstrate the efficacy of these ML approaches in predicting CS, with statistical
analyses showcasing strong correlations between predicted and experimental outcomes,
particularly with boosting achieving high accuracy levels. K-fold cross-validation further
validates model accuracy, emphasizing the superiority of boosting over ANN in terms of
error minimization.

Beskopylny, A. N., Stel’makh, S. A., Shcherban’, E. M., Mailyan, L. R.,

Meskhi, B., Razveeva, I., ... & Beskopylny, N. (2022). Concrete strength
prediction using machine learning methods CatBoost, k-nearest
neighbors, support vector regression. Applied Sciences, 12(21), 10864.

The paper investigates the utility of machine learning techniques, specifically CatBoost
gradient boosting, k-nearest neighbors, and support vector regression (SVR), in
predicting the compressive strength of concrete (Beskopylny et al., 2022). It underscores
the growing interest in employing artificial intelligence algorithms to enhance concrete
production processes, with a particular emphasis on predicting concrete properties
accurately. By leveraging empirical data accumulated from laboratory experiments on
 self-compacting concretes, the study aims to develop and compare the performance of
these machine learning algorithms. Through a thorough literature review, the paper
contextualizes the significance of accurate predictions in the construction industry, where
the mechanical properties of building materials like concrete play a pivotal role in
ensuring structural integrity and performance.

The methodology section outlines the approach taken to build and evaluate the machine
learning models. It details the parameters selected for each algorithm, the dataset used for
training and testing, and the performance evaluation metrics employed (Beskopylny et
al., 2022). The study adopts a rigorous process of model building and training, including
parameter tuning using grid search and cross-validation techniques. Furthermore, it
discusses the significance of parallelizing the optimization process to reduce
computational time. By providing a detailed overview of the experimental setup and
model development process, the paper lays the foundation for a comprehensive analysis
of the predictive capabilities of the machine learning algorithms considered.

Chen, H., Li, X., Wu, Y., Zuo, L., Lu, M., & Zhou, Y. (2022).
Compressive strength prediction of high-strength concrete using long
short-term memory and machine learning algorithms. Buildings, 12(3),
302.
The prediction of compressive strength of high-strength concrete (HSC) is critical in the
construction industry due to its importance in structural integrity and performance. Traditional
methods for testing compressive strength are often uneconomical, time-consuming, and labor-
intensive. To address this, machine learning (ML) algorithms have been increasingly explored
for their predictive capabilities. Chen et al. (2022) examined the effectiveness of Long Short-
Term Memory (LSTM) networks and Support Vector Regression (SVR) in predicting the
compressive strength of HSC. Their study utilized a dataset comprising 324 samples with five
input variables: water, cement, fine aggregate, coarse aggregate, and superplasticizer. The study
aimed to compare the performance of LSTM and SVR using metrics such as root mean square
error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and the
correlation coefficient (R²).

The literature indicates that LSTM networks, introduced by Hochreiter and Schmidhuber in
1997, are well-suited for time-series prediction tasks due to their ability to handle long-term
dependencies and mitigate issues such as gradient disappearance and explosion (Hochreiter &
Schmidhuber, 1997). LSTM's architecture includes forget, input, update, and output gates, which
collectively enhance its predictive performance. In the context of HSC, LSTM has shown
superior accuracy compared to traditional methods. Similarly, SVR, an application of Support
Vector Machines (SVM) in regression problems, offers advantages in handling nonlinear
regression and obtaining global optimal solutions (Smola & Schölkopf, 2004). The use of kernel
functions in SVR allows for mapping data to higher-dimensional spaces, enhancing prediction
accuracy. The comparative analysis in Chen et al.'s study revealed that LSTM significantly
outperformed SVR, with an R² of 0.997 and lower error metrics, making it a more reliable tool
for pre-estimating HSC compressive strength before laboratory tests. This underscores the
potential of advanced ML algorithms in optimizing construction processes and ensuring
structural safety (Chen et al., 2022).

Ahmad, A., Ahmad, W., Chaiyasarn, K., Ostrowski, K. A., Aslam, F.,
Zajdel, P., & Joyklad, P. (2021). Prediction of geopolymer concrete
compressive strength using novel machine learning
algorithms. Polymers, 13(19), 3389.
Geopolymer concrete (GPC) has garnered attention for its sustainability and lower
environmental impact compared to conventional concrete, as it uses industrial by-products like
fly ash and slag, significantly reducing CO2 emissions (Davidovits, 2008). The application of
machine learning (ML) in civil engineering has further enhanced the predictive capabilities for
material properties, with studies demonstrating the effectiveness of algorithms such as artificial
neural networks (ANNs) and boosting methods in accurately forecasting compressive strength
(Yeh, 1998; Dinh et al., 2018). ANN models, inspired by the human brain's neural structure, are
effective in handling non-linear relationships and large datasets (Haykin, 2009). Meanwhile,
boosting algorithms, including AdaBoost, combine multiple weak learners to form a strong
predictor, offering superior accuracy in predictions (Freund & Schapire, 1997). These
advancements underscore the potential of combining GPC with advanced ML techniques to
optimize material performance and sustainability in construction.

Paudel, S., Pudasaini, A., Shrestha, R. K., & Kharel, E. (2023).

Compressive strength of concrete material using machine learning
techniques. Cleaner Engineering and Technology, 15, 100661.
The study under review builds on the extensive body of research focused on improving the
compressive strength (CS) of concrete by incorporating fly ash (FA) as a partial cement
replacement. Given the complex factors influencing concrete strength, such as aggregate shape
and water-cement ratios, predicting CS accurately remains challenging. The literature highlights
various machine learning (ML) approaches to tackle this challenge, with specific emphasis on
non-ensemble models like Multiple Linear Regression (MLR) and Support Vector Regression
(SVR), as well as ensemble models like AdaBoost, Random Forest, XGBoost, and Bagging.
Previous studies have utilized these algorithms with datasets of varying sizes and compositions
to enhance predictive accuracy. Key findings indicate that ensemble models, particularly
XGBoost, outperform non-ensemble models due to their ability to handle complex interactions
among input parameters. The reviewed study aims to compare these ML models using a robust
dataset of 633 experimental results, assessing model performance with metrics such as MAE,
MSE, R², MAPE, RMSE, and the a20-index. Notably, XGBoost emerged as the most accurate,
achieving a high coefficient of determination (R² = 0.95) and low error margins, while MLR was
the least accurate. Sensitivity analysis further underscored the significance of curing days,
cement, and water content in determining CS, affirming the potential of advanced ML techniques
in optimizing concrete mix designs economically and efficiently, without extensive experimental
trials.

Asteris, P. G., Koopialipoor, M., Armaghani, D. J., Kotsonis, E. A., &

Lourenço, P. B. (2021). Prediction of cement-based mortars compressive
strength using machine learning techniques. Neural Computing and
Applications, 33(19), 13089-13121.
Machine learning (ML) techniques are crucial for predicting the compressive strength of cement-
based mortars. Support Vector Machines (SVM) use kernel functions to map data into high-
dimensional spaces, with key parameters like regularization (C) and kernel width (γ). Decision
Trees (DT) split data into zones using decision rules, requiring pruning to avoid overfitting.
AdaBoost enhances weak classifiers by iteratively adjusting instance weights, minimizing errors
and preventing overfitting. Each method offers unique benefits and challenges, making them
valuable for modeling material properties.

References
1. Al-Shamiri, A. K., Kim, J. H., Yuan, T. F., & Yoon, Y. S. (2019). Modeling the compressive
strength of high-strength concrete: An extreme learning approach. Construction and Building
Materials, 208, 204-219.
2. Ahmad, W., Ahmad, A., Ostrowski, K. A., Aslam, F., Joyklad, P., & Zajdel, P. (2021). Application
of advanced machine learning approaches to predict the compressive strength of concrete
containing supplementary cementitious materials. Materials, 14(19), 5762.
3. Song, H., Ahmad, A., Farooq, F., Ostrowski, K. A., Maślak, M., Czarnecki, S., & Aslam, F. (2021).
Predicting the compressive strength of concrete with fly ash admixture using machine learning
algorithms. Construction and Building Materials, 308, 125021.
4. Cao, R., Fang, Z., Jin, M., & Shang, Y. (2022). Application of machine learning approaches to
predict the strength property of geopolymer concrete. Materials, 15(7), 2400.
5. Moein, M. M., Saradar, A., Rahmati, K., Mousavinejad, S. H. G., Bristow, J., Aramali, V., &
Karakouzian, M. (2023). Predictive models for concrete properties using machine learning and
deep learning approaches: A review. Journal of Building Engineering, 63, 105444.
6. Ahmad, A., Ahmad, W., Chaiyasarn, K., Ostrowski, K. A., Aslam, F., Zajdel, P., & Joyklad, P.
(2021). Prediction of geopolymer concrete compressive strength using novel machine learning
algorithms. Polymers, 13(19), 3389.
7. Beskopylny, A. N., Stel’makh, S. A., Shcherban’, E. M., Mailyan, L. R., Meskhi, B., Razveeva, I.,
& Beskopylny, N. (2022). Concrete strength prediction using machine learning methods CatBoost,
k-nearest neighbors, support vector regression. Applied Sciences, 12(21), 10864.
8. Chen, H., Li, X., Wu, Y., Zuo, L., Lu, M., & Zhou, Y. (2022). Compressive
strength prediction of high-strength concrete using long short-term memory and
machine learning algorithms. Buildings, 12(3), 302.
9. Ahmad, A., Ahmad, W., Chaiyasarn, K., Ostrowski, K. A., Aslam, F., Zajdel, P., &
Joyklad, P. (2021). Prediction of geopolymer concrete compressive strength using
novel machine learning algorithms. Polymers, 13(19), 3389.
10. Ahmad, A., Ahmad, W., Chaiyasarn, K., Ostrowski, K. A., Aslam, F., Zajdel, P., &
Joyklad, P. (2021). Prediction of geopolymer concrete compressive strength using
novel machine learning algorithms. Polymers, 13(19), 3389.
11. Paudel, S., Pudasaini, A., Shrestha, R. K., & Kharel, E. (2023). Compressive
strength of concrete material using machine learning techniques. Cleaner
Engineering and Technology, 15, 100661.
12. Asteris, P. G., Koopialipoor, M., Armaghani, D. J., Kotsonis, E. A., & Lourenço, P. B. (2021).
Prediction of cement-based mortars compressive strength using machine learning
techniques. Neural Computing and Applications, 33(19), 13089-13121.
13. Ghafari, E., Bandarabadi, M., Costa, H., & Júlio, E. (2015). Prediction of fresh
and hardened state properties of UHPC: comparative study of statistical mixture
design and an artificial neural network model. Journal of Materials in Civil
Engineering, 27(11), 04015017.
14. Kumar, A., Arora, H. C., Kapoor, N. R., Mohammed, M. A., Kumar, K.,
Majumdar, A., & Thinnukool, O. (2022). Compressive Strength Prediction of
Lightweight Concrete: Machine Learning Models. Sustainability 2022, 14, 2404.
15.