CIOT-BASED EARLY DIAGNOSIS OF HEART FAILURE FROM MULTIMODAL DATA USING CHI-SQUARE-BASED DEEP NEURAL CLASSIFIER

A. K. S. Saranya and T. Jaya

Keywords

IoT, cloud storage, MIT-BIH arrhythmia repository, medical records, RPA, RPA loss function, Chi-square-based deep neural network

Abstract

A significant number of human lives can be saved by monitoring heart patients effectively based on heart signals. For patients and doctors, the classification and forecasting of heart diseases based on electrocardiogram (ECG) signals have become increasingly important over the past ten years. In this paper, for monitoring, forecasting, and heartbeat diagnostics, a new mobile healthcare application built on the Cloud and Internet of things (CIoT) has been developed. A healthcare system built on the CIoT consists of three components. The first section uses IoT devices, the MIT-BIH arrhythmia repository, and medical records to gather the necessary data. The second section is used to store medical records on a cloud database safely. The heartbeat detection prediction is done in the third portion. The Robotic process automation (RPA) learning component retrieves the useful features, normalises the feature values, and then does estimating using the RPA loss function. Finally, the diagnosis of the heartbeat detection using the Chi-square- based deep neural network (CSDNN) is analysed. Results were analysed to show how well the suggested methodology performed when compared to other deep learning methods, like convolutional neural network, long short-term memory (CNN- LSTM), contextual online learning under local differential privacy (LCOL), and coy- grey wolf optimisation-based deep convolution neural network (coy-GWO-CNN).

Important Links:



Go Back