Prediction of Cardiac Disease at Early Stage Using Feature Selection Based Tuned-Support Vector Machine

Main Article Content

A. Sankari Karthiga, Dr. M. Safish Mary

Abstract

A cardiac disease prediction model is presented in this work. Smart Device-enabled healthcare is a critical component of contemporary technology. Medical sensors are used in healthcare to continuously give a large number of medical data. In the healthcare industry, data is generated at a rapid pace, resulting in a large amount of data. Heart disease diagnosis has been discovered to be a difficult issue that can provide a computerised estimate of the severity of heart disease, allowing for easier follow-up action. As a result, the diagnosis of heart disease has attracted a lot of interest in the medical community around the world. The use of feature selection algorithms in the detection of heart disease was found to be quite effective. In this paper, we look at how Support Vector Machine performance is affected by feature selection algorithms including information gain and correlation features. In addition, we proposed the Tuned Support Vector Machine as a novel model. The Tuned-experimental Support Vector Machine's (Tuned-SVM) results are compared to those of other feature selection algorithms like Correlation Feature Selection and Information Gain. To evaluate Tuned Support Vector Machine's good performance, performance measurements are used. The proposed framework is demonstrated in the PYTHON environment with a dataset collected from UCI heart disease database.

Article Details

Section
Articles