Optimizing Diabetes Diagnosis Through Pulse Waveform Analysis and Data Mining.

Area of Science:

• Medical Informatics
• Machine Learning in Healthcare
• Computational Biology

Background:

• Diabetes mellitus poses a significant global health challenge.
• Accurate and early diagnosis is crucial for effective diabetes management.
• Existing diagnostic methods can be enhanced through advanced computational approaches.

Purpose of the Study:

• To develop a robust diabetes diagnosis model using ensemble learning techniques.
• To evaluate the performance of a Stacking classifier with Multilayer Perceptron as the final model.
• To identify the most effective diagnostic model for diabetes detection.

Main Methods:

• Utilized four weak learners: Random Forest, Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Decision Tree algorithms.
• Applied a Stacking ensemble method to combine weak learners.
• Selected Multilayer Perceptron (MLP) as the meta-classifier for the final model.
• Evaluated model performance using a dataset from Changhua Christian Hospital, Taiwan.

Main Results:

• The Stacking ensemble learning method significantly improved model performance.
• The optimized model achieved an F1 score of 0.86.
• The optimized model demonstrated an Area Under the Curve (AUC) score of 0.90.

Conclusions:

• The proposed Stacking ensemble model is effective for diabetes diagnosis.
• The optimized model shows high accuracy and discriminative power.
• This approach offers a promising tool for enhancing diabetes detection in clinical settings.