Symptom-Based COVID-19 Prognosis through AI-Based IoT: A Bioinformatics Approach

Area of Science:

• Healthcare technology
• Machine learning applications
• Predictive analytics in medicine

Background:

• The conventional healthcare approach faces challenges in disease prediction and remote patient monitoring.
• Internet of Things (IoT) and machine learning offer a transition to patient-specific healthcare solutions.
• IoT enables real-time data collection and smarter decision-making to enhance healthcare quality.

Purpose of the Study:

• To develop an automated IoT-based system for COVID-19 prognosis using machine learning models.
• To compare the performance of various machine learning classifiers for symptom-based disease prediction.
• To establish a framework for real-time and remote healthcare monitoring.

Main Methods:

• Comparative analysis of machine learning classifiers including logistics regression, k-nearest neighbor, support vector machine, random forest, decision trees, Naïve Bayes, and gradient booster.
• Utilizing cloud-stored data for model validation and verification.
• Measuring model performance based on accuracy for COVID-19 prognosis.

Main Results:

• The k-nearest neighbor (k-NN) model demonstrated the highest accuracy at 97.97% for COVID-19 prognosis.
• Decision tree (97.79%) and support vector machine (97.42%) also showed high accuracy.
• Other models like logistics regression, random forest, gradient boosting, and Naïve Bayes had varying levels of accuracy.

Conclusions:

• The developed IoT framework has significant clinical value for real-time and remote healthcare monitoring.
• The findings can assist global healthcare systems in managing COVID-19 and future pandemics.
• Machine learning models, particularly k-NN, show strong potential for accurate symptom-based disease prognosis.