Machine Learning Based Diabetes Classification and Prediction for Healthcare Applications.

Area of Science:

• Biomedical Informatics
• Machine Learning in Healthcare
• Public Health Data Analysis

Background:

• Advancements in biotechnology and healthcare generate vast amounts of sensitive patient data.
• Intelligent data analysis aids in early disease detection and prevention, crucial for managing life-threatening conditions like diabetes mellitus.
• Diabetes significantly increases the risk of secondary complications such as heart, kidney, and nerve damage.

Purpose of the Study:

• To propose a machine learning-based approach for the classification, early identification, and prediction of diabetes.
• To present a hypothetical Internet of Things (IoT)-based system for continuous blood glucose (BG) monitoring.
• To evaluate the efficacy of different machine learning models for diabetes-related tasks.

Main Methods:

• Employed Random Forest (RF), Multilayer Perceptron (MLP), and Logistic Regression (LR) for diabetes classification.
• Utilized Long Short-Term Memory (LSTM), Moving Averages (MA), and Linear Regression (LR) for predictive analysis.
• Validated the models using the benchmark PIMA Indian Diabetes dataset.

Main Results:

• The Multilayer Perceptron (MLP) classifier achieved an accuracy of 86.08% for diabetes classification.
• Long Short-Term Memory (LSTM) demonstrated improved predictive accuracy for diabetes, reaching 87.26%.
• Comparative analysis showed the proposed approach's adaptability for public healthcare applications.

Conclusions:

• Machine learning models, particularly MLP and LSTM, show significant promise for accurate diabetes classification and prediction.
• The proposed methods offer a viable approach for early diabetes identification and management.
• The integration of IoT for blood glucose monitoring can enhance patient care and disease management strategies.