Disease Prediction for Heart using Data Mining Techniques and Decision Tree Classification

Abstract

Heart disease is a prominent cause of death worldwide, and many people are concerned about it.  Early detection of heart disease becomes very crucial and has the potential to save many lives. However, detecting cardiovascular diseases such as heart attacks, coronary artery disease, and others of similar types is a critical challenge presented by routine clinical data analysis, even by using well known algorithms. We can’t risk lives of people so the result or prediction should be correct up to large extent. Machine learning (ML) can provide an efficient solution for making decisions and making accurate predictions in such crucial medical scenarios. The suggested study uses a dataset as well as data mining approaches such as various categorization algorithms throughout the report. Machine learning employs techniques such as Random Forest and Decision Tree. The innovative technique for the machine learning model is being developed here. The implementation employs three machine learning algorithms: The K Nearest Neighbors (KNN) classifier, the Decision Tree classifier, and the Random Forest classifier are the three types of classifiers. Here, decision tree classification is a subpart of random forest or we can say that random forest is built by several decision trees. These three machine learning techniques falls under supervised learning, where we work on labeled data and also feedback is provided to the model. The interface is intended to collect the user’s input parameters in order to predict heart disease with the help of the various attributes that are related to heart disease, for which we used a KNN, Decision Tree and Random Forest model. The concept of heat maps is used to show correlation among various features and using that important decisions can be concluded. The concept of optimum k value is also employed to get better classification results by observing score of the model via that.