ENHANCED SALP SWARM OPTIMIZATION WITH ARTIFICIAL NEURAL NETWORK TRAINING FOR HEART MONITORING IN IOMT CLOUD ARCHITECTURE

Abstract

The use of Internet of Medical Things (IoMT) in human-healthcare systems has enabled the gathering of sensor data for the diagnosis and prognosis of cardiac disease. However, handling big data, including clinical, omics, and health data, in actual- time can be challenging due to issues such as noise, size, formats, missing values, and a large number of features. This can make it difficult for health monitoring systems to gather accurate data. To address these challenges, an consideration based Convolution Neural Network methodology (ACNN) and an Enhanced Salp Swarm Optimization established Extended Short-Term Memory technique (Enhanced SSO-LSTM) model have been proposed. This approach consists of four layers: the actual-data gathering level, cloud data - storage layer, data analytics layer, and display layer. The primary sources of data handled by the data gathering layer include wearable sensor devices, medical records, the heart dataset, and the diabetes dataset. The cloud storage layer uses a wireless network to store all the data that has been gathered from the cloud server and various datasets. In the data analytics layer, large data analytics operations such as pre-processing, dimensionality reduction, filtering, and classification are carried out. A Health care expert can evaluate the patient's condition using the Enhanced SSO-LSTM classification findings in the display layer. The Enhanced SSO-LSTM model achieved accuracy in the range 96% with accuracy 93%, recollection of 85%, and F-measure in 80% of ranges. The ACNN model had 93% accuracy, 89% precision, 89% recall, and 79% F-measure. Overall, the Enhanced SSO- LSTM model demonstrated improved performance in large dataset health monitoring.