Optimised Heart Disease Prediction Scheme

Abstract

Cardio Vascular Disease (CVD) includes heart, peripheral arterial and coronary heart disease as well as cerebrovascular disease. The symptoms of this disease include fast heartbeats, pain or discomfort in the centre of the chest, dizziness, difficult breathing, overweight, hypertension and high cholesterol. To determine accurate and efficient treatment of heart disease, specialized cardiologists must conduct complicated tests and procedures. Early detection is the key to preventing CVD. Difficulties in diagnosing CVD and transporting patients over long distances lead to an unjustified increase in death rate which is a burden on patients. Among important risk factors for cardiovascular problems are an abnormal build-up of fats and cholesterol in the blood vessels, harmful alcohol consumption, smoking and a lack of proactive steps, and poor dietary habits. Early detection and cardiovascular management are essential factors to lowering the incidence of CVD. Applying an optimized technique would be a better choice for such classification. The proposed system uses Modified Teaching Learning Based Optimization (MTLBO), Kernel Density Function (KDF) and Density-based Modified Teaching Learning Based Optimization (DMTLBO) to perform dataset classification. Classification is performed using Support Vector Machine (SVM), Ensemble Learning (EL-Adaboosting). It is seen that DMTLBO_Adaboosting offers better results based on Accuracy, Precision, Recall, Time Period, F-Measure as well as Error Rate.