Certain investigations on heart disease prediction using hybrid classifier models through machine learning

Abstract

Globally, Congenital Heart disease is the seventh most common reason for infant deaths (Baranwal et al. 2023). Based on Heart Disease and Stroke Statistics 2008, in United states almost 30% of infant mortality occurred in 2004 as a result of congenital cardiac disease. According to Sameni et al. (2010), one out of every 125 neonates is born with a cardiac defect each year. On May 8, 2013, The Hindu, a daily newspaper in India, reported that daily, almost three lakh infants pass away in the time period of 24 hours after they are born. Early detection of cardiac defects and regular foetal heart monitoring can help pediatric cardiologists address these issues and support them in taking preventative measures throughout gestation. The majority of cardiac disorders have symptoms in the shape of the cardiac electrical signal. The noninvasive technique of monitoring cardiac signals is advised for tracking foetal health issues. Fetal ElectroCardioGram (FECG) signal acquired from the mother s abdomen using the noninvasive acquisition technique has greater artifacts, with Mother s noise being the main source. The extraction of FECG is difficult because there is a significant mother s component in the recorded abdominal signal. Therefore, a technique to totally eliminate this component must be devised in order to obtain the desired FECG. newline