Deep Transfer Learning for Communicable Disease Detection and Recommendation in Edge Networks

Area of Science:

• Medical Informatics
• Machine Learning
• Wearable Technology

Background:

• Rising global communicable disease cases necessitate rapid, reliable detection methods.
• Limitations in PCR test availability and performance hinder early outbreak control.
• Resource-constrained edge devices pose challenges for complex data analysis in healthcare.

Purpose of the Study:

• To develop an edge-centric healthcare framework for timely communicable disease detection.
• To integrate wearable sensors and advanced machine learning for efficient data analysis.
• To address the challenges of analyzing data on resource-limited edge devices.

Main Methods:

• Implementation of an edge-centric healthcare framework with wearable sensors.
• Preprocessing of sensor-collected features into a usable dataset.
• Application of Deep Transfer Learning (DTL) for efficient data analysis on edge devices.

Main Results:

• The proposed DTL technique demonstrated superior performance compared to existing methods.
• Achieved a high accuracy of 99.8% in disease prediction.
• Validated on a benchmark COVID-19 dataset with 11 features and 2 million sensor data points.

Conclusions:

• The developed DTL-integrated framework offers an efficient and accurate solution for early disease detection.
• This approach overcomes the limitations of traditional diagnostic methods and resource constraints of edge devices.
• Enables timely decision-making for communicable disease management, minimizing delays.