EARLY DETECTION OF CARDIAC ARREST IN NEONATES USING MACHINE LEARNING AND STATISTICAL MODELS
Keywords:
Neonatal Cardiac Arrest , Machine Learning , Statistical Models , Neonatal Intensive Care Unit (NICU) , Logistic Regression, Support Vector Machine, Artificial Neural Network , Early DetectionAbstract
Neonatal cardiac arrest is a critical medical emergency that requires prompt recognition and intervention to reduce mortality risk and improve the infant's health outcomes. Conventional monitoring techniques can inadequately identify early warning indicators in neonatal intensive care units (NICUs). This research advocates for the use of physiological and clinical parameters, including as respiratory rate, heart rate, oxygen saturation, birth weight, and gestational age, to swiftly detect newborns undergoing cardiac arrest. The methodology is based on statistical models and machine learning techniques. To improve prediction accuracy, various machine learning algorithms are analyzed. The ensemble comprises Naïve Bayes, Artificial Neural Networks (ANN), Decision Tree Classifier, Support Vector Machine (SVM), and Logistic Regression. Statistical preprocessing and feature extraction are utilized to improve model efficacy and decrease the incidence of erroneous predictions. A research revealed that the suggested prediction technique effectively identifies infants at risk of cardiac arrest. Furthermore, it is accurate and uniform. The proposed strategy may enable early intervention and enhance the survival rate of infants, ultimately benefiting physicians and nurses in NICUs.