| dc.description.abstract | The Quantum Support Vector Machine (QSVM) is a hybrid machine learning approach that combines the computational power of quantum computing with classical classification algorithms. This study aims to evaluate the performance of QSVM in classifying asphalt mixture quality based on Marshall Stability (MS) and Marshall Flow (MF) values. The dataset used consists of technical parameters of the mixture, including asphalt content (Pb), specific gravity (Gmb, Gmm), air voids (Va), voids in mineral aggregate (VMA, VFA), as well as aggregate characteristics (Gsb, Gse). The MS and MF targets were divided into two classes (low and high) using the quantile binarization method. The QSVM model was constructed using a quantum kernel based on ZZFeatureMap and executed through statevector simulation on a classical backend. Evaluation results indicate that QSVM achieved an accuracy of 90.2% for MS prediction and 86.9% for MF, with macro F1-scores of 0.88 and 0.83, respectively. These results demonstrate the superiority of QSVM in binary classification compared to conventional regression models such as Artificial Neural Network (ANN), Random Forest (RF), Multi Expression Programming (MEP), and XGBoost. This classification approach is also considered more practical for field applications, particularly in the context of rapid decision-making regarding asphalt mixture quality assessment. The study recommends QSVM as an intelligent alternative for data-driven asphalt quality control systems. | en_US |
