Innovative AI Approaches for Concrete Strength Prediction: Towards Sustainable Buildings

Concrete's compressive strength is crucial for sustainability and quality. Advanced machine learning methods can minimize environmental impact and enhance mix design. Hybrid and ensemble-based approaches are being explored for accurate strength estimates, promoting sustainable construction techniques in civil engineering. ML techniques provide accurate first predictions of desired results. To predict concrete compressive strength, the study assesses several machine learning techniques, including Random Forest (RF), Decision Trees (DT), Multi-layer Perceptron (MLP), and Bagging. This study used open-source ML software called Waikato Environment for Knowledge Analysis (Weka). The dataset was obtained from the Kaggle platform, which was divided into a training dataset and a testing dataset. The features of this dataset were fed to the model. Then, the model was evaluated to examine the effectiveness of the suggested model in predicting concrete strength. This assessment used statistical indicators that include coefficient correlation (R²), mean absolute error (MAE), relative absolute error (RAE), mean square error (MSE), and root mean square error (RMSE). According to the experimental data results, RF had the highest coefficient correlation, at 0.9604. However, the coefficient correlation for Bagging was 0.9384, the DT was 0.9238, and the coefficient correlation for MLP was 0.8722.