Lifestyle Disease Influencing Attribute Prediction Using Novel Majority Voting Feature Selection

Abstract

In humans, Lifestyle Disease (LSD) is caused by an improper way of life such as less physical activity, sleeplessness, unhealthy eating habits, liquor drinking, and smoking. LSD leads to gastric problems, indigestion of food, and prognosis to heart problems, Type II diabetes, and lung diseases. LSD treatment and medication lead to high expenditure for patients and country through LSD management and policies. Patients who suffer from LSD need lifelong treatment. The solution to reducing mortality due to Lifestyle Diseases is early detection and effective treatment. LSDs are low progressive in nature and need an effective and accurate early prediction method for effective treatment. The most prevalent LSD, based on World Health Organization (WHO) statistics, are heart disease and diabetes problems. This proposed model identifies the influencing attributes for contributing disease risk such as diabetes and heart attack and their associations using novel feature selection techniques such as Novel Majority Voting Ensembled Feature Selection (NMVEFS) for heart disease (HD) and diabetes. The influencing attributes are used to build a Clinical Decision Support System (CDSS) for LSD using a deep neural network, which helps physicians in identifying heart disease and diabetes at an early stage with 97.5% prediction accuracy and decreases treatment cost.