Neonatal Health Prediction System for IoT based Smart Incubator Parameters

Abstract

Neonatal disease prediction remains a critical challenge in healthcare, requiring advanced models to improve early diagnosis and intervention. This study presents NeoCoD (Neonatal Cause of Disease Predictor), a novel deep learning-based model designed to enhance predictive accuracy for neonatal diseases. NeoCoD integrates transfer learning through pre-trained BERT for feature extraction and Bidirectional LSTM networks to capture temporal dependencies in sequential data. The model demonstrates superior performance compared to existing methods, achieving 92.58% accuracy, 91.80% precision, and 92.10% recall. Its advanced deep learning techniques and robust generalization capabilities offer significant improvements in neonatal disease prediction. NeoCoD's performance highlights its potential as a valuable tool for early diagnosis and intervention, addressing critical needs in neonatal healthcare.