An Analytic Approach to Diagnose Heart Stroke Using Supervised Machine Learning Techniques

• Published in: Intelligent Healthcare pp. 133 - 162
• Main Authors: Majumdar, Anurima, Roy, Sunipa, Chakraborty, Chinmay
• Format: Book Chapter
• Language: English
• Published: Singapore Springer Nature Singapore 2022
• Subjects: AUC-ROC plot; Confusion matrix; Decision trees; Intelligent health management systems (IHMS); K-nearest neighbors; Logistic regression; MI prediction; Naïve Bayes; Random forest; Supervised machine learning; Support vector machine
• ISBN: 981168149X; 9789811681493
• DOI: 10.1007/978-981-16-8150-9_7

This chapter proposed how machine-learning algorithms can be used in developing intelligent health management systems (IHMS) specially for cardiovascular disease. With the increasing load of cardiovascular disease and events on the human race the proper diagnosis of the disease within time and prediction of the disease can save more lives. In this chapter the rudiments of machine learning algorithms are discussed. Six Supervised classification based techniques are analysed and compared to find the most suitable model to predict cardiovascular disease or chances of cardiovascular events for the provided dataframe. Six algorithms are used here. The Confusion matrix for each of the classifiers is evaluated. Comparative bar plot and AUC ROC plot is generated to measure the executions of the prototypes. It is observed that Support-Vector-Machine and K-Nearest-Neighbors achieved the highest accuracy (96.20%) which makes them most suitable for heart attack or myocardial infarction (MI) prediction for this given dataset.