Quantum Variational Based Support Vector Machine for Early Detection of Sepsis
Rana Veer Samara Sihman Bharattej R, Ramy Riad Al–Fatlawy, S Meenakshi Sundaram, E. Annie Rathnakumari, M. Sudha · 2024
The sepsis is a very toxic disorder with associated clinical appearance which is difficult to identify and treat. The early analysis and proper treatment are challenging to minimize the death and encourage subsistence in the suspected cases which enhance the detection result. The screening prediction systems that estimating and evaluating the early detection of patient weakness because the efficacy is imperfect at the distinct level. To overcome these challenges, the Machine Learning (ML) techniques is used for predicting sepsis in early phases. In this research the Quantum Variational based Support Vector Machine (QV-SVM) is proposed for early detection of sepsis. It improves scalability and efficiency through leveraging quantum speedups, permitting quick procedure of large datasets. The min-max normalization approaches scale the information to an explicit range, enhancing the performance through confirming features contribute equally. The proposed QV-SVM achieves accuracy of 98.25%, recall of 96.58%, precision of 97.59% and F1-score of 97.08 % thorough Medical Information Mart for Intensive Care (MIMIC) III dataset as compared to exiting methods Extreme Gradient Boosting (XG Boost).