Implementation of Machine Learning Algorithm for Cardiac Arrest Prediction

Machine learning (ML) is a subfield of AI that uses statistical algorithms. Cardiac Arrest or heart failure has been implicated as one of the leading causes of death. The limited accuracy and the inherent invasiveness in diagnosis of this disease call for a revamp of the existing diagnostic protocol.  In this study, we developed Machine learning (ML) algorithms for the prediction of cardiac arrest. Our protocol employs different methods for classification of the HD dataset using univariate and Bivariate analysis for prediction of cardiac arrest on input data which contains 11 features such as ChestPainType, age, gender etc and Pair plot to check the distribution of each variable and how it correlated with the target variable (Cardiac Arrest). Our result indicated that the ASY pain type was the highest ChestPainType that had cardiac arrest with 54% while NAP had 22%, ATA had 19% and TA 5%. The male genders were also observed to have the highest rate of cardiac arrest when compared to the female genders. Our protocol was able to predict the occurrence of cardiac arrest and at the same time recommend possible treatments, medication and exercises regime to the patient via the web application interface.