Heart failure prediction using Pegasos Quantum Support Vector Classifier

Abstract

Heart failure (HF) is a critical global health problem, and early diagnosis with accuracy is highly essential for effective treatment. the combination of machine learning (ML) with quantum physics has driven research into quantum algorithms for data processing, offering new methods in novel computational environments. While quantum computing is still in its infancy, with quantum algorithms being actively tested and developed, this study proposes a Quantum Machine Learning (QML) approach for HF prediction using the Pegasos Quantum Support Vector Classifier (QSVC) algorithm. Addressing a critical research gap—the lack of direct comparisons between this algorithm and classical ML algorithms in HF — this study conducts a comprehensive evaluation of Pegasos (Primal Estimated sub-GrAdient SOlver for SVM) QSVC against classical models, including Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), and k-Nearest Neighbors (K-NN), using a real-world electronic health dataset. The proposed work utilizes quantum feature maps and kernels to improve HF prediction. the results showed that Pegasos QSVC leveraging quantum principles has potential in handling high dimensional data. The results highlight complementary strengths of the quantum and classical approaches. This work establishes the foundation for further QML research highlighting the technologies potential to transform healthcare predictive modeling.