Parametric Analysis on Heart Disease Prediction using Ensemble based Classification

Abstract

Heart disease is a bigger cause of morbidities and mortalities amongst the populace of the world. Predicting of heart disease is observed as a vital subject in clinical data investigation. The quantity of data in medical industry is massive. Data mining turned the larger compilation of medical data to information for making knowledgeable predictions and decisions. This research work develops a new heart disease prediction model that includes 3 most important phases viz. "proposed feature extraction, dimensionality reduction and proposed ensemble based classification". At first, the higher and lower statistical features are extracted. Nevertheless, due to "curse of dimensionality" it is necessary to reduce the extracted features, for which "Principal Component Analysis (PCA) is employed. These features are then classified via "Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbour (KNN) with optimized Neural Network (NN)". Moreover, the training of NN is done by S-CDF via fine-tuning the optimal weights. Finally, parametric analysis is held to confirm the effectiveness of the developed model.