Enhanced IoT-Based Health Data Fusion with Recursive Feature Elimination for Improved Patient Monitoring

Abstract

Healthcare market utilization of IoT relies on digital systems that monitor and analyze health issues within the framework of IoT technology. Through IoT and smart devices a high level of smart environment can be built. Medical devices linked to smartphone apps enable the collection of medical health data and other necessary patient information. Data Fusion (DF) functions as a process of uniting information that exists across multiple data sources. These methods get implemented beyond text processing applications within other domains. DF operates on numerous distributed data sources to minimize detection errors as well as increase their reliability in multisensory settings. The main goal targets scalability together with performance enhancement and identification capabilities. The environmental responsiveness of medical devices depends on their strength to adjust output according to varying circumstances. The system demonstrates its scalability through expected function with continuous performance while utilizing available resource management at optimal levels. Forming a specialized group to establish common practices among tracking devices remains vital because this ensures uniformity in communications and data processing and user interface standards. The primary research outputs include implementing pre-processing followed by Improvised Context-aware Data Fusion (ICDF), Improved Principal Component Analysis (IPCA) for feature extraction and Enhanced Recursive Feature Elimination (ERFE) for selection as well as ensemble-based Machine Learning (ML) model as classifier. The Improved Dynamic Bayesian Network (IDBN) presents itself as a balanced choice for tractability since it can serve as a tool for ICDF operations. Results from simulations indicate that the proposed ICDF model performs best in the healthcare system with 96% accuracy along with 95% precision and 96% recall and F1 score at 96%.