Machine Learning Based Prognosis of Heart Malady

Abstract

Heart disease is a complex condition that affects a large number of people around the globe. In healthcare particularly in field of cardiology, accurate diagnosis of cardiac disease is critical. We offer a system for identifying cardiac illness that is both efficient and accurate which is based on the machine learning techniques. Among the various classification algorithms like, Support vector machine, K-nearest neighbor, Artificial neural network, Logistic regression, Decision tree and Naive Bayes we used Random Forest algorithm in this project, whereas traditional feature selection algorithms such as Break, for reducing unnecessary and superfluous features, we used the least absolute shrinkage selection operator, as well as local learning and minimal redundancy. To overcome the feature selection complication, we suggested a new fast conditional mutual information feature selection approach. The feature selection techniques are used to boost the classification accuracy and lower the classification system's execution time. In addition, the leave one subject out cross validation technique was utilized to discover best practices in hyper parameter tweaking and model assessment. The accuracy score measures are used to evaluate the classifiers' performance. The performance of the classifiers was evaluated using features chosen using feature selection methods. Furthermore, the suggested approach may be simply used in healthcare to detect cardiac problems.