An Enhanced Fusion Approach for Meticulous Presaging of HD Detection Using Deep Learning

Abstract

Numerous methodologies have been adopted for the detection of arrhythmia. ECG signals are challenging to utilize customary diagnosis of arrhythmia by the use of machine learning. It is capable observing of heart arrhythmia suffered patients could save their life. In the early stages have improved the results of patients to detect the disease. Arrhythmias is a condition where the electrical signal of the heart is detected either quicker, more slow than typical, this is a reason for death for people each year. In this current study for the detection of HD, the NCA with LSTM is used. The CA dataset has taken from UCI, which consists 279 characteristics. The objective is to characterize CA patients into 16 classes by ECG dataset for getting prediction output. The DL method as (LSTM) and LDA is used. NCA is implemented on both this model for selecting the relevant features from the arrhythmia dataset by the combination of LSTM +NCA and NCA+LDA obtained better results in terms of accuracy rate that is 98.6 % and 94.1 respectively.