Predicting Sepsis in the Intensive Care Unit (ICU) through Vital Signs using Support Vector Machine (SVM)
Zeina Rayan, Marco Alfonse, Abdel-Badeeh M. Salem
Abstract
Background: As sepsis is one of the life-threatening diseases, predicting sepsis with high accuracy could help save lives. Methods: Efficiency and accuracy of predicting sepsis can be enhanced through optimal feature selection. In this work, a support vector machine model is proposed to automatically predict a patient’s risk of sepsis based on physiological data collected from the ICU. Results: The support vector machine algorithm that uses the extracted features has a great impact on sepsis prediction, which yields the accuracy of 0.73. Conclusion: Predicting sepsis can be accurately performed using the main vital signs and support vector machine.
Topics & Concepts
Support vector machineSepsisIntensive care unitComputer scienceArtificial intelligenceVital signsRelevance vector machineMachine learningFeature (linguistics)Feature selectionIntensive care medicinePattern recognition (psychology)MedicineImmunologySurgeryLinguisticsPhilosophySepsis Diagnosis and TreatmentMachine Learning in HealthcareTraditional Chinese Medicine Studies