Diabetes diagnosis using a hybrid CNN LSTM MLP ensemble

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Computational Biology

Background:

• Diabetes mellitus is a serious chronic condition requiring timely diagnosis for effective management.
• Traditional diabetes diagnosis methods relying on clinical and physical data are often time-consuming.
• There is a need for automated, efficient, and accurate diagnostic tools for diabetes.

Purpose of the Study:

• To develop and evaluate an ensemble deep learning model for automated diabetes diagnosis.
• To improve the efficiency and accuracy of diabetes identification compared to conventional methods.
• To leverage deep learning techniques for analyzing complex patient data.

Main Methods:

• Data preprocessing involved cleaning, normalization, and organization for deep learning models.
• Feature extraction utilized a Convolutional Neural Network (CNN) for spatial characteristics and a Long Short-Term Memory (LSTM) network for temporal data.
• An ensemble approach combined CNN and LSTM features, feeding them into a Multi-layer Perceptron (MLP) classifier for final diagnosis.

Main Results:

• The proposed ensemble deep learning model demonstrated superior performance over existing methods.
• The model achieved an average accuracy of 98.28% in diagnosing diabetes.
• The system reached a precision of 0.99%, indicating high reliability in its diagnostic predictions.

Conclusions:

• The developed ensemble deep learning model offers a highly accurate and efficient automated solution for diabetes diagnosis.
• This approach shows significant potential in improving early detection rates and patient outcomes.
• The integration of CNN and LSTM networks provides a robust framework for analyzing medical data for disease diagnosis.