Various Feature Selection Techniques Used for Predicting and Diagnosing Heart Disease

Abstract

To effectively treat cardiac patients before a heart attack occurs, a precise prognosis of heart disease is necessary. Recently, machine learning-based algorithms for predicting and diagnosing heart disease have been described. However, the lack of a sophisticated framework that can use several sources of data to forecast cardiac disease means that current algorithms cannot manage large datasets. These systems use standard methods for selecting data points and assigning weights to them according to their relevance. Heart disease diagnosis has also failed to benefit from the use of these techniques. A review of the various feature selection methods used in the detection of heart disease is provided in this chapter of the book. The data used in the trials comes from a UCI library and relates to heart disease. In order to test the biomedical system's efficiency, many well-known validation methods have been used. This allows doctors to recognize heart disease in patients at an early stage so that more treatment can be started.