Stroke Disease Prediction based on ECG Signals using Deep Learning Techniques

Abstract

Stroke-related diseases are rapidly increasing day by day due to the changes in environmental factors including lifestyles, food habits, and stress-related working cultures. According to a recent report from World Health Organization (WHO), Stroke is the second largest disease after cardiovascular disease that leads to death. Early diagnosis of stroke-related diseases was one of the major requirements for patients as well as medical professionals. Deep learning techniques are one of those methods that are suitable for stroke disease diagnosis when deployed properly. This research proposed a medical framework approach for the detection of abnormalities in the ECG data related to stroke diseases. ECG plays a vital role in the detection of several stroke risk factors including left ventricular hypertrophy. This work proposed a framework based on Long Short-term Memory (LSTM) network for predicting stroke-related diseases with ECG data and other parameters. The experimental results show that 90% accuracy results with the combination of ECG data and the Deep learning approach. Finally, Receiver operating characteristic (ROC) curves has shown promising results. This work also proved that the model is suitable for the early detection of stroke-related diseases with minimum overhead in terms of efficiency.