Cardiac Arrhythmia Classification Using Neural Networks with Selected Features

Abstract

This research is to present a new approach for cardiac arrhythmia disease classification. An early and accurate detection of arrhythmia is highly solicited for augmenting survivability. In this connection, intelligent automated decision support systems have been attempted with varying accuracies tested on UCI arrhythmia data base. One of the attempted tools in this context is neural network for classification. For better classification accuracy, various feature selection techniques have been deployed as prerequisite. This work attempts correlatio n-based feature selection (CFS) with linear forward selection search. For classification, we use incremental back propagation neural network (IBPLN), and Levenberg-Marquardt (LM) classification tested on UCI data base. We compare classification results in terms of classification accuracy, specificity, sensitivity and AUC. The experimental results presented in this paper show that up to 87.71% testing classification accuracy can be obtained using the average of 100 simulations.