Exploring the Effectiveness of Computational Methods in Detecting Heart Disease

Formation of a waxy material in the coronary arteries is a sign of heart disease. If nothing else, this buildup of plague-like wax in the arteries slows down the flow of blood and, in the worst-case scenario, results in death. To predict cardiac disease, this work uses learning models based on k-nearest neighbour (KNN) and support vector machine (SVM). The effectiveness of the suggested model is assessed by an experiment. Furthermore examined is the outcome of the experimental assessment of the suggested model's predictive capability. The South Africa Heart Disease Dataset is utilized to gather information for the research on heart disease. The source is the South African Coronary Risk Factor Study carried out in 1983 by Rousseauw and colleagues. For the existence of a myocardial infarction, the dataset contains a binary answer. The target audience is men between the ages of 15 and 64. It has 462 observations in total. In myocardial infarction, 160 of these findings are consistent. Not a single myocardial infarction is seen in 302 observations. There are several factor-level and numerical predictors in the dataset. Lastly, a test set analysis is conducted to evaluate the KNN and SVM algorithm performance. Following performance examination, the KNN's accuracy is 64.51% and the SVM's accuracy is 73.91% based on experimental findings on the South African heart disease data pool.