An enhanced diabetes prediction amidst COVID-19 using ensemble models

Area of Science:

• Medical Informatics
• Computational Biology
• Public Health

Background:

• Diabetes prediction is crucial for patient outcomes, especially post-COVID-19.
• Machine learning offers advanced tools for early disease detection.
• Existing methods require optimization for improved accuracy.

Purpose of the Study:

• To enhance the accuracy of diabetes prediction using machine learning.
• To explore feature engineering and ensemble methods for better predictive performance.
• To provide a robust model for early diabetes diagnosis and management.

Main Methods:

• Utilized machine learning classifiers, including Support Vector Machines (SVM).
• Applied advanced feature engineering (interaction, polynomial features).
• Developed an ensemble model integrating Decision Trees, Gradient Boosting, and SVM.

Main Results:

• Initial SVM model achieved 76.62% accuracy.
• Feature engineering identified significant correlations (e.g., Glucose).
• The ensemble model reached a high accuracy of 93.2%.

Conclusions:

• Ensemble machine learning models significantly improve diabetes prediction accuracy.
• Advanced feature engineering enhances model performance by uncovering data patterns.
• This approach offers a blueprint for improved early diagnosis, personalized treatment, and preventive care.