Efficient Parameter Optimization of Quantum Support Vector Classifier Using Ant Colony Optimization for Medical Data Classification

Healthcare systems are instrumental in augmenting human health, thereby necessitating the continuous development of highly efficacious models. With the rapid expansion of these models across diverse medical disciplines, the potential for innovative applications continues to rise at an exponential rate. Nevertheless, developing prompt, precise, and efficient models tailored for medical use cases presents a substantial hurdle. Addressing this issue, a new medical data classification model leverages the Quantum Support Vector Classifier (QSVC) and ant colony optimization (ACO). The proposed modified version of ACO is employed to find the optimal hyperparameter tuning values of QSVC. The optimization procedure concentrated on adjusting the QSVC’s hyperparameters, particularly the count of circuit repetitions and the nature of entanglement found in the quantum circuit of the QSVC. Three medical benchmark datasets collected from the UCI machine learning repository are adopted. The experimental results demonstrated that the proposed model is very promising compared to the other machine learning algorithms and state-of-the-art models. It obtained an overall accuracy of 100% for the Breast Cancer Wisconsin Diagnostic dataset, 91.62% for the Heart Attack Analysis Prediction dataset, and 80.35% for the Diabetes dataset.