Machine Learning for Online Automatic Prediction of Common Disease Attributes Using Never-Ending Image Learner
E Rajesh1, Shajahan Basheer1, Rajesh Kumar Dhanaraj1
1School of Computing Science and Engineering, Galgotias University, Greater Noida 203201, India.
Diagnostics (Basel, Switzerland)
|January 8, 2023
Summary
Automatic online prediction (OAP) uses machine learning and a Never-Ending Image Learner to accurately predict disease attributes from images. This improves upon existing methods for online healthcare diagnostics.
Area of Science:
- Computer Science
- Artificial Intelligence
- Medical Informatics
Background:
- Online healthcare systems are growing, but user-submitted health queries on forums lack reliability and accuracy.
- Mild symptoms often deter individuals from seeking in-person medical consultations, leading to reliance on unverified online information.
- Existing online health prediction methods face challenges with accuracy and user response.
Purpose of the Study:
- To propose and evaluate an Automatic Online Prediction (OAP) system for disease attribute prediction.
- To leverage machine learning, specifically the Never-Ending Image Learner, for enhanced online healthcare diagnostics.
- To improve the accuracy and efficiency of predicting common disease attributes from limited image data.
Main Methods:
- Developed an Automatic Online Prediction (OAP) system using a Never-Ending Image Learner (NEIL) with machine learning.
- Employed a multi-access edge computing platform for machine-learning-assisted prediction from multiple images via multiple-instance learning.
- Utilized M-theory for efficient real-time image prediction and isotropic positioning for image data storage.
Main Results:
- The proposed OAP method demonstrated higher accuracy in predicting common disease attributes compared to existing approaches.
- The system achieved improved operating efficiency through the machine learning of multiple images with isotropic positioning.
- Performance metrics confirmed the superior accuracy and efficiency of the NEIL-based OAP system.
Conclusions:
- The proposed machine learning-based Automatic Online Prediction system offers a more accurate and efficient solution for online disease attribute prediction.
- The integration of Never-Ending Image Learner and isotropic positioning enhances the reliability of online healthcare diagnostics.
- This approach addresses the limitations of traditional online health forums by providing dependable, AI-driven medical insights.
