Novel Internet of Things based approach toward diabetes prediction using deep learning models

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Internet of Things (IoT)

Background:

• Technological advancements enable the integration of IoT and machine learning (ML) in healthcare.
• Medical IoT creates interconnected environments for improved patient care and early disease detection.
• Early disease forecasting aids medical professionals in timely interventions.

Purpose of the Study:

• To develop a smart patient health monitoring system using ML for early and accurate chronic disease detection.
• To utilize IoT sensors for patient data capture and ML for analysis and prediction of diseases like diabetes.

Main Methods:

• Implementation involved using a diabetic dataset.
• Six ML techniques were employed: Support Vector Machine (SVM), Logistic Regression, Artificial Neural Network (ANN), Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), and Long Short-Term Memory (LSTM).
• Performance was evaluated using accuracy, precision, recall, and F1-Score.

Main Results:

• Recurrent Neural Network (RNN) demonstrated superior performance with 81% accuracy, 75% precision, and 65% F1-Score.
• Artificial Neural Network (ANN) achieved the highest recall at 56% compared to other models.
• The proposed system enables earlier diagnosis of chronic diseases.

Conclusions:

• The developed smart patient health monitoring system effectively utilizes IoT and ML for early disease detection.
• RNN and ANN show promise for accurate diabetes prediction, aiding clinical decision-making.
• This system can significantly improve patient outcomes through timely diagnosis and intervention.