Heart Disease Prediction using Different Machine Learning Algorithms

Abstract

Identifying a person's potential for developing heart disease is one of the most challenging tasks medical professionals faces today. With nearly one death from heart disease every minute, it is the leading cause of death in the modern era [4]. The database is taken from Kaggle. Various machine learning algorithms are used for prediction of heart disease detection here are Random Forest, XG-Boost, K- Nearest Neighbors (KNN), Logistic Regression, Support Vector Machines (SVM). All these algorithms are implemented using Python programming with Google collab.  The performance evaluation parameters used here are Accuracy, precision, recall and Fi-score. Training and testing are implemented for different ratios such as 60:40, 70:30 and 80:20. From the analysis and comparisons of evaluation parameters of all the above algorithms, XG-Boost is having the highest accuracy and recall value. KNN having worst accuracy and recall amongst all. XG-Boost is having a training accuracy of 98.86, 98.74 and 97.68 for training and testing ratio of 60:40, 70:30 and 80:20 respectively. XG-Boost is having a testing accuracy of 95.85, 95.45 and 96.09 for training and testing ratio of 60:40, 70:30 and 80:20 respectively. So, XG-Boost algorithm can be used for obtaining the best prediction for heart disease.  This type of heart disease prediction can be used as a secondary diagnostic tool for doctors, for best and fast prediction. This can help the early prediction of heart disease thus increasing the chances of the saving the life heart patient.