Analysis of The Diagnostic Parameters of Heart Diseases and Prediction of Heart Attacks

Abstract

Medical data is made up of a huge number of heterogeneous variables gathered from various sources all of which provide a different perspective on a patient's condition. Machine Learning proves to be very effective method for the prediction of unstructured data. Algorithms such as SVC, K Nearest Neighbor, Random Forest Classifier, Naïve Bayes etc. can be used for the early detection for the disease. Data mining technique are used to gather data from health care databases and are used for making clinical decision of the disease at the preliminary level without the intervention of the medical experts.[1] Using the state-of-the-art wearable electronic equipment can also be used for collecting continuous data from the patients. The classification techniques in the area of Machine Learning in the medical field, with the goal to find similar patterns, thereby producing vital predictions, and being useful in early diagnosis of the disease is the focus of this research paper. The algorithm which fits the data and predicts with more accuracy is analyzed. The novelty in this research is predicting if a patient already with a heart disease will get a heart attack or not. Whereas, most of the researchers are interested only in predicting the presence of a heart disease. This paper focuses on the prediction of heart attacks in patients having a heart disease.