Diabetes Mellitus Prediction and Severity Level Estimation Using OWDANN Algorithm.

Area of Science:

• Biomedical Informatics
• Machine Learning in Healthcare
• Computational Biology

Background:

• Diabetes mellitus is a prevalent chronic disease with serious complications.
• Accurate early diagnosis is crucial for reducing diabetes risk and incidence.
• Existing prediction methods often suffer from low accuracy and high training times.

Purpose of the Study:

• To propose an Optimized Whale Decomposition Artificial Neural Network (OWDANN) algorithm for diabetes mellitus prediction.
• To develop a system for estimating the severity level of diabetes.
• To address the limitations of existing methods in terms of accuracy and training efficiency.

Main Methods:

• The proposed system involves two phases: disease prediction and severity level estimation.
• For prediction, the Pima dataset undergoes preprocessing, feature extraction, and classification using OWDANN.
• For severity estimation, a diabetes-positive dataset is preprocessed, features are extracted, and severity is predicted using GDHC.

Main Results:

• The proposed system achieved high performance metrics: 98.97% accuracy, 94.98% sensitivity, 95.62% specificity, 97.02% precision, 93.84% recall, and 94.04% F-measure.
• The system demonstrated low False Discovery Rate (FDR) at 0.094% and False Positive Rate (FPR) at 0.023%.
• The OWDANN algorithm significantly outperformed existing state-of-the-art methods in diabetes prediction and severity estimation.

Conclusions:

• The developed OWDANN algorithm offers a highly accurate and efficient solution for diabetes mellitus prediction.
• The system's ability to estimate severity levels provides valuable clinical insights.
• This approach holds significant potential for improving early detection and management of diabetes.