Published Paper Details:

We are using Machine learning (ML) to predict concrete strength using various input parameters. This study uses a large dataset from several construction projects and various ML algorithms, such as multiple linear regression (ML), decision trees (ML), random forests (ML), support vector machines (SVM), and artificial neural networks (ANN). The results demonstrate high accuracy levels with the ANN model having the highest performance. The implications of this study for the construction industry are to streamline the design process, reduce reliance on laboratory testing and optimize material usage while meeting safety standards and reducing costs. However, there are limitations, such as dataset size, real world conditions, and fine-tuning. Hyperparameter tuning uses random search to train models with higher predictive performance. The missing data is filled with the average of the data available, allowing for more data to be included in the training. Comparative studies on two well-known compressive datasets of tensile strengths and high performance concrete suggest that the current approach is significantly improved in terms of prediction accuracy as well as computational effort. According to comparative studies, for this specific prediction problem, GBR and xGBoost-trained models perform better than models based on SRV and MLP