Predicting Heart Disease Using Gaussian Confidence Distance Algorithm with Extra Tree Classifier

To prevent fatal heart failure early detection of Heart attack is essential steps. Heart Disease prediction happened well in advance using Machine learning, Artificial algorithms in Information Communication Technology, and the use of a combination of multiple algorithms can further improve accuracy of predictions. However, it is important to note that the accuracy of algorithms in predicting heart disease depends on the quality and quantity factors. Important privacy concerns and potential biases are also very much essential in data. Reason for heart disease includes improper food habits and an increase in fat content due to less physical work. The diagnosis of heart illness is mainly related to symptoms, and clinical examination of patients. In this study, we try to classify and predict heart diseases at early stages on specifical features by fusing Various Machine Learning Algorithms. Machine learning when being used in health-care is able to detect disease earlier and accurately. This work yielded an accuracy of 88.5% in predicting heart disease using Random Forest Algorithm. To increase the prediction, many algorithms were combined with more tree classifiers to select features with domain-knowledge adding and physician expertise to the prediction models. This can lead to more accurate and reliable predictions, that will ultimately improve patient diagnosis and reduce health-care costs.