RFFE - Random Forest Fuzzy Entropy for the classification of Diabetes Mellitus

Area of Science:

• Medical Informatics
• Machine Learning in Healthcare
• Computational Biology

Background:

• Diabetes mellitus is a chronic metabolic disorder characterized by elevated blood glucose levels due to insufficient insulin production.
• Complications of diabetes affect vital organs like the retina, kidneys, and nerves, necessitating lifelong management.
• Early detection and risk assessment are critical for preventing diabetes and its associated health issues.

Purpose of the Study:

• To develop and evaluate a novel prototype for early diabetes prediction using machine learning.
• To enhance diabetes prediction accuracy by integrating Fuzzy Entropy with Random Forest algorithms.
• To compare the performance of the proposed model against various established machine learning techniques.

Main Methods:

• The study employed a prototype incorporating data imputation, sampling, and feature selection.
• Key prediction techniques included Fuzzy Entropy, Synthetic Minority Oversampling Technique (SMOTE), Convolutional Neural Network (CNN) with Stochastic Gradient Descent with Momentum (SGDM), Support Vector Machines (SVM), Classification and Regression Tree (CART), K-Nearest Neighbor (KNN), and Naïve Bayes (NB).
• The Pima Indian Diabetes (PID) dataset was utilized for model training and validation, with performance assessed via confusion matrix and ROCAUC.

Main Results:

• The proposed Random Forest Fuzzy Entropy (RFFE) model demonstrated superior performance in diabetes prediction.
• The RFFE model achieved a high accuracy rate of 98% on the PID dataset.
• Comparative analysis confirmed the effectiveness of RFFE over other tested machine learning algorithms.

Conclusions:

• The Random Forest Fuzzy Entropy (RFFE) approach is a highly effective and valuable tool for early diabetes prediction.
• Accurate early detection through advanced machine learning models can significantly aid in diabetes prevention and management strategies.
• The findings highlight the potential of integrating Fuzzy Entropy with ensemble methods for improved chronic disease risk assessment.