A Method to Improve Human Heart Disease Prediction Using Machine Learning Algorithms

Abstract

There is great diversity in the field of medical science due to computational power and technical innovation, especially in identifying human heart disease. Today it is one of the deadliest human heart diseases in the world and have very serious effects on human life. Accurate and timely identification of heart disease in humans can be very helpful in prevent heart failure in its early stages and will improve patient survival. Manual method for determining the heart disease is biased and can vary between researchers. In this regard, efficient and reliable machine learning algorithms resources for detecting and classifying people with heart disease and those who are healthy. According to suggestion in our study, we identified and predicted heart disease in humans using a variety of machine learning algorithms and using heart disease dataset to evaluate its performance using various measures, such as sensitivity, specificity, F-measure, and classifier accuracy. For this purpose, we used nine machine learning classifiers for the final dataset before and after hyper parameter tuning of machine learning classifiers, such as AB, LR, ET, MNB, CART, SVM, LDA, RF, and XGB. In addition, we verify their accuracy on a standard heart disease dataset by performing several standardized, pre-processing procedures of the data set and hyper parameter tuning. In addition, to train and validate machine learning algorithms, we implemented standard K-fold cross-validation technique. Finally, the experimental results show that the accuracy of the predictive classifiers with improved hyper parameter tuning and achieved remarkable results with data normalization and hyper parameter tuning of machine learning classification.