An Efficient IoT-based Patient Monitoring System for Heart Disease Prediction and Categorization using Optimized Cell Attention based Spiking Neural Network

• Published in: International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (Online) pp. 47 - 54
• Main Authors: Suryawanshi, P B, Rawat, Ruchira, Deepa, Battu, Chakraborty, Subhra, Mohanty, Aparajita, Taqui, Syed Noeman
• Format: Conference Proceeding
• Language: English
• Published: IEEE 03.10.2024
• Subjects: Biomedical monitoring; Cell Attention Spiking Neural Network (CASNN); Diseases; Heart; Hybrid Mud Ring Osprey Optimization Algorithm (HMROOA); Internet of Things; Internet of Things (IoTs); Kaggle Dataset; Medical services; Optimization; Patient monitoring; Prediction algorithms; Predictive models; Smart Healthcare; Spiking neural networks; War Strategy Optimization (WSO); Wearable Sensors
• ISSN: 2768-0673
• DOI: 10.1109/I-SMAC61858.2024.10714815

The smart healthcare patient monitoring system based on Internet of Things (IoTs) capabilities continuously gathers physiological information through wearable sensors attached to the person's body. In this concern, the proposed study presents a novel smart healthcare patient monitoring system based on Cell Attention Spiking Neural Network (CASNN). The proposed CASNN model is obtained by integrating Cell Attention Network (CAN) with Spiking Neural Network (SNN) that provides enhanced performance during heart disease prediction and classification. Initially, preprocessing is performed to remove unwanted noises and duplicate values from the raw physiological data. Secondly, selection of optimal feature subsets is carried out by utilizing Hybrid Mud Ring Osprey Optimization Algorithm (HMROOA). Finally, cardiac disease prediction and categorization is accomplished with the help of CASNN, in which the hyper parameters are optimized using War Strategy Optimization (WSO). As a result, the proposed model accurately identifies the presence or absence of heart disease, which helps in timely diagnosis of heart illness. The physiological data used for analysis is taken from UCI Heart Disease Kaggle Dataset in which 75% of data is used for training and 25% of data are used for testing. The investigational results show that the developed system achieves greater detection accurateness of 99.7%, which is better than the conventional heart disease prediction models.