A combined approach WNN for ECG feature based disease classification

Abstract

ECG observes the movements in the heart by sensing the electrical variations in the human skin by using
small sensors which are implanted on the chest of the patients. In traditional work many algorithms have been developed
for detecting heart disease from which it is concluded that most of the work in biomedical ECG analysis for the
prediction or detection was done only up-to the feature extraction, no further decision system development field was
highlighted. So, in the proposed work, after getting the information from the ECG, dataset standalone system has
developed in this paper. Considering this fact, a novel approach has identified termed as Aritifical neural network where
the dataset is trained and then classified in order to detect the type of disease a person is suffering from. Specifically,
proposed system works on two types of diseases such as Bradycardia and Tachycardia. Apart from this, it also classified
either the person is normal or not and if not then which type of disease he or she has. The proposed system is compared
with the traditional approach using MATLAB software tool and performance analysis are carried out in the end of the
paper. The evaluation of simulation analysis confirmed that proposed system is more efficient and accurate in
comparison with the traditional wavelet based approach in terms of overall success rate and error rate of the system