The present work provides a detailed analysis of machine learning methods which can successfully be used in discovering and predicting coronary heart disease from datasets containing human body parameters. Unsupervised learning techniques are responsible for detecting patterns linking together medical parameters which result in a common outcome, and our paper details three such approaches: clustering (GMM, hierarchical), outlier detection (KDE, KNN density, KNN-ARD) and association mining. At the same time, supervised learning classifiers will help in understanding the prediction possibilities of such a medical condition, and this has been done by applying and statistically evaluating the performance of logistic regression and artificial neural networks (ANN) against a predictive baseline model. The results of our analysis show that predicting coronary heart disease is a well-suited task for machine learning detection, with small error ranges being achieved even from low-dimension datasets of only a dozen features, provided that the features are relevant and the patient data is well-recorded.
Software And Hardware
• Hardware: Processor: i3 ,i5 RAM: 4GB Hard disk: 16 GB • Software: operating System : Windws2000/XP/7/8/10 Anaconda,jupyter,spyder,flask Frontend :-python Backend:- MYSQL