An Optimized Support Vector Machine-Based Embedded Framework for Real-Time Personalized Health

Abstract

Continuous monitoring of physiological parameters is essential for early detection of health risks. This paper proposes an optimized embedded health monitoring system using Particle Swarm Optimization-based Support Vector Machine (PSO-SVM). The system collects real-time data such as heart rate, blood pressure, glucose level, and temperature using multiple sensors. A dataset of 300 samples is used for training and evaluation. The data is preprocessed using noise filtering and normalization techniques. Statistical features are extracted to represent physiological behavior. PSO is used to optimize SVM parameters for improved classification. The proposed model achieves an accuracy of 97.1% with high precision and recall. It outperforms conventional SVM (94.5%) and Random Forest (96.2%). The system effectively detects abnormal health conditions and is suitable for real-time healthcare applications.