| Research on utilizing wastewater in concrete to reduce freshwater consumption is increasing.
However, wastewater impacts concrete strength and accurately predicting compressive strength is
essential for improving design, construction, safety, performance, durability, and sustainability.
While experimental approaches to explore this possibility are time-consuming, labor-intensive, and
costly, this study leverages machine learning techniques to predict the compressive strength of
concrete containing wastewater and to identify the most effective model. A dataset of 706 samples
from past studies on concrete with wastewater was used to train and test eight machine learning
models: k-nearest neighbor (KNN), support vector regression (SVR), decision tree (DT), random
forest (RF), adaptive boosting (AdaBoost), gradient boosting regression (GBR), extreme gradient
boosting (XGBoost), and categorical boosting (CatBoost). The dataset was preprocessed and split
into an 80-20 training-test set, with hyperparameter optimization performed using grid search and 10-
fold cross-validation. Model performance was evaluated using mean absolute error (MAE), root mean
squared error (RMSE), mean absolute percentage error (MAPE), and coefficient of determination
(R²). The input parameters, including the pH value of wastewater, cement content, water content,
water-cement ratio, fine and coarse aggregate contents, percentage of wastewater, and age of testing,
were used to predict the concrete compressive strength. The results showed that ensemble models,
particularly CatBoost, GBR, and XGBoost, outperformed simpler models like KNN and SVR in
predictive performance and generalization ability. CatBoost emerged as the top performer with the
highest R² value of 0.964 and the lowest error metrics (MAE of 1.4 MPa, RMSE of 1.979 MPa,
MAPE of 5.29%) on the test set. Additionally, areas for future research are identified, including
exploring diverse wastewater types, incorporating additional concrete properties, and enhancing
model robustness in real-world conditions.
Keywords: Concrete; Wastewater; Sustainability; Compressive Strength; Machine Learning |