Cardiac disease detection and analysis – a Machine Learning based approach
Sheikh Afaan Farooq, Ajatray Swagat Bhuyan, Richa Sharma, Hamnah Rao, Mohammad Qamar, Shuvendu Das
Abstract
The cardiac disease can be denoted as a general disorder that includes ischemic heart disorder, cardiac infarction and Arrhythmia, each being deadly and a cause of about 1.20 crore deaths as revealed in a disturbing report compiled by the World Health Organization. This research paper proposes an algorithm to diagnose cardiac diseases in a person while making use of the fast machine learning based classifier model built on top of decision tree algorithm which is selected as being the most accurate among a group of 5 algorithms. It also detects the most significant symptoms related to cardiac diseases by calculating the importances of features utilized in the dataset. In short, this research proposes a fast and simple cardiac diseases classifier which can detect the disease with an accuracy of 93.3%, utilizing the UCI based heart disease dataset which is an improvement over the previous studies done and models developed on the same dataset. The model can then also accept responses or new data which can be fed to the classifier for classifying heart disease on the new data.