<![CDATA[This paper proposes a hybrid predictive model that integrates Support Vector Regression (SVR) with a Genetic Algorithm (GA) to estimate the compressive strength of concrete using slump test data, thereby offering an alternative to conventional, resource-intensive laboratory testing. The employed dataset encapsulates the nonlinear relationship between concrete slump and compressive strength. Given the sensitivity of SVR to hyperparameter selection, its standalone application yielded suboptimal predictive performance. To mitigate this, GA was utilized for hyperparameter optimization, selected for its effectiveness in global search and handling complex parameter spaces compared to traditional optimization techniques. The SVR-GA model was systematically evaluated against established machine learning algorithms, including Decision Tree, Neural Network, Naïve Bayes, and K-Nearest Neighbors, chosen based on their prevalence and diverse methodological characteristics. Performance evaluation incorporated robust validation methods to prevent overfitting and ensure generalizability. The results indicate that the proposed model delivers rapid and accurate predictions, suitable for practical, on-site application, with the potential to significantly reduce time and costs associated with traditional testing. Limitations related to dataset specificity and model generalizability are acknowledged. Future research directions include extending the framework to additional concrete properties and the development of real-time predictive systems. A schematic representation of the SVR-GA integration is included.]]>
Copyrights © 2024
