<abstract xmlns="http://www.w3.org/1999/xhtml">

<p>Millions of people worldwide suffer from diabetes, a medical condition that is spreading at an accelerating pace. Numerous studies show that risk factors that may arise from diabetes can be avoided if the disease is detected early. The health-care monitoring system has benefited greatly from early diabetes prediction made possible by the integration of Deep Learning (DL) and Machine Learning (ML) algorithms. The objective of many early studies was to increase the prediction model accuracy. However, DL algorithms often cannot fully exploit the potential of the available datasets because they are too small. This study includes a very accurate DL model as well as a novel system that integrates cloud services and allows users to directly enhance an existing data set, which can increase the accuracy of DL techniques. Therefore, the Long Short-Term Memory (LSTM) model with controller is chosen for efficient type-1 diabetes prediction. Experimental validation of the proposed Nonlinear Model Predictive Control (NMPC)_LSTM algorithm method is compared with other conventional DL algorithms. The proposed controller method achieves excellent blood glucose set point tracking and the proposed algorithms achieves 98.95% accuracy for the obtained data. It outperforms other existing methods with an increase in percentage accuracy compared to the Benchmark Pima Indian Diabetes Datasets (PIDD).</p>
</abstract>

Millions of people worldwide suffer from diabetes, a medical condition that is spreading at an accelerating pace. Numerous studies show that risk factors that may arise from diabetes can be avoided if the disease is detected early. The health-care monitoring system has benefited greatly from early diabetes prediction made possible by the integration of Deep Learning (DL) and Machine Learning (ML) algorithms. The objective of many early studies was to increase the prediction model accuracy. However, DL algorithms often cannot fully exploit the potential of the available datasets because they are too small. This study includes a very accurate DL model as well as a novel system that integrates cloud services and allows users to directly enhance an existing data set, which can increase the accuracy of DL techniques. Therefore, the Long Short-Term Memory (LSTM) model with controller is chosen for efficient type-1 diabetes prediction. Experimental validation of the proposed Nonlinear Model Predictive Control (NMPC)_LSTM algorithm method is compared with other conventional DL algorithms. The proposed controller method achieves excellent blood glucose set point tracking and the proposed algorithms achieves 98.95% accuracy for the obtained data. It outperforms other existing methods with an increase in percentage accuracy compared to the Benchmark Pima Indian Diabetes Datasets (PIDD).

A Cloud-Connected Digital System for Type-1 Diabetes Prediction using Time Series LSTM Model

Department of Electronics and Communication Engineering

Department of Information Systems, College of Computer and Information Sciences

Princess Nourah Bint Abdulrahman University (PNU)

Applied College at Mahayil, King Khalid University

Measurement Science Review

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Millions of people worldwide suffer from diabetes, a medical condition that is spreading at an accelerating pace. Numerous studies show that risk factors that...

A Cloud-Connected Digital System for Type-1 Diabetes Prediction using Time Series LSTM Model

Publicado en línea: 13 abr 2024

Páginas: 83 - 87

Recibido: 18 sept 2023

Aceptado: 12 mar 2024

DOI: https://doi.org/10.2478/msr-2024-0011

Palabras clave
Cloud controller, diabetes, long short-term memory, glucose-insulin, non-linear predictor, time series

© 2024 K. Priyadarshini et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

A Cloud-Connected Digital System for Type-1 Diabetes Prediction using Time Series LSTM Model

Publicado en línea: 13 abr 2024

Páginas: 83 - 87

Recibido: 18 sept 2023

Aceptado: 12 mar 2024

DOI: https://doi.org/10.2478/msr-2024-0011

Palabras claveCloud controller, diabetes, long short-term memory, glucose-insulin, non-linear predictor, time series

© 2024 K. Priyadarshini et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Palabras clave
Cloud controller, diabetes, long short-term memory, glucose-insulin, non-linear predictor, time series