Cardiac Disease Analysis Using Machine Learning

Abstract

Goal of this paper is to analyse the accuracies of all algorithms and determine which algorithm is best for predicting heart disease in humans. We took a consideration amount of data and trained it using python for better performance and identification of the effected part of the heart for every algorithm. We use Python because it is very good and supportive for data visualisation and understanding complex patterns. Recently, there are some of the diseases which cannot be predicted by the existing system so due to which many medical universities are conducting challenges to solve them and predict such type of diseases with machine learning, for example aorta separation and atrial tubes overflow are some of the challenges held. Because the algorithm can only track a limited number of patterns, and because new problems might arise at any time, it is not always possible to predict the outcome of cardiovascular disease accurately. Pre-processing the data is the initial phase in the machine learning process, proceeded with feature extraction based on the data cleaning, classifying, and performance review. The outcome's accuracy is improved with XGBoost.