Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Comparative Study
. 2024 Sep 19;14(1):21863.
doi: 10.1038/s41598-024-70277-x.

Data-driven blood glucose level prediction in type 1 diabetes: a comprehensive comparative analysis

Hoda Nemat  1 , Heydar Khadem  2 , Jackie Elliott  3   4 , Mohammed Benaissa  2
Affiliations
Comparative Study

Data-driven blood glucose level prediction in type 1 diabetes: a comprehensive comparative analysis

Hoda Nemat et al. Sci Rep. .

Abstract

Accurate prediction of blood glucose level (BGL) has proven to be an effective way to help in type 1 diabetes management. The choice of input, along with the fundamental choice of model structure, is an existing challenge in BGL prediction. Investigating the performance of different data-driven time series forecasting approaches with different inputs for BGL prediction is beneficial in advancing BGL prediction performance. Limited work has been made in this regard, which has resulted in different conclusions. This paper performs a comprehensive investigation of different data-driven time series forecasting approaches using different inputs. To do so, BGL prediction is comparatively investigated from two perspectives; the model's approach and the model's input. First, we compare the performance of BGL prediction using different data-driven time series forecasting approaches, including classical time series forecasting, traditional machine learning, and deep neural networks. Secondly, for each prediction approach, univariate input, using BGL data only, is compared to a multivariate input, using data on carbohydrate intake, injected bolus insulin, and physical activity in addition to BGL data. The investigation is performed on two publicly available Ohio datasets. Regression-based and clinical-based metrics along with statistical analyses are performed for evaluation and comparison purposes. The outcomes show that the traditional machine learning model is the fastest model to train and has the best BGL prediction performance especially when using multivariate input. Also, results show that simply adding extra variables does not necessarily improve BGL prediction performance significantly, and data fusion approaches may be required to effectively leverage other variables' information.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1
Figure 1
A schematic diagram demonstrating the preprocessing steps.
Figure 2
Figure 2
The colour-coded surveillance error grid related to the predictions of CTF approach with univariate input 30 min in advance for patient 570.
Figure 3
Figure 3
The colour-coded surveillance error grid related to the predictions of TML approach with univariate input 30 min in advance for patient 570.
Figure 4
Figure 4
The colour-coded surveillance error grid related to the predictions of DNN approach with univariate input 30 min in advance for patient 570.
Figure 5
Figure 5
The colour-coded surveillance error grid related to the predictions of CTF approach with multivariate input 30 min in advance for patient 570.
Figure 6
Figure 6
The colour-coded surveillance error grid related to the predictions of TML approach with multivariate input 30 min in advance for patient 570.
Figure 7
Figure 7
The colour-coded surveillance error grid related to the predictions of DNN approach with multivariate input 30 min in advance for patient 570.
Figure 8
Figure 8
CD diagram of comparing different prediction models with univariate input pairwisely over the data contributors of Ohio_2018 dataset for the 60-minute prediction horizon based on RMSE metric.
Figure 9
Figure 9
CD diagram of comparing different prediction models with univariate input pairwisely over the data contributors of Ohio_2018 dataset for the 60-min prediction horizon based on MAE metric.
Figure 10
Figure 10
CD diagram of comparing different prediction models with univariate input pairwisely over the data contributors of Ohio_2018 dataset for the 60-minute prediction horizon based on SE metric.
Figure 11
Figure 11
CD diagram of comparing different prediction models with multivariate input pairwisely over the data contributors of Ohio_2018 dataset for the 30-min prediction horizon based on RMSE metric.
Figure 12
Figure 12
CD diagram of comparing different prediction models with multivariate input pairwisely over the data contributors of Ohio_2018 dataset for the 60-min prediction horizon based on MAE metric.
Figure 13
Figure 13
CD diagram of comparing different prediction models with multivariate input pairwisely over the data contributors of Ohio_2018 dataset for the 30-min prediction horizon based on SE metric.
Figure 14
Figure 14
CD diagram of comparing different prediction models with multivariate input pairwisely over the data contributors of Ohio_2018 dataset for the 60-min prediction horizon based on RMSE metric.
Figure 15
Figure 15
CD diagram of comparing different prediction models with multivariate input pairwisely over the data contributors of Ohio_2018 dataset for the 60-min prediction horizon based on MAE metric.
Figure 16
Figure 16
CD diagram of comparing different prediction models with multivariate input pairwisely over the data contributors of Ohio_2018 dataset for the 60-min prediction horizon based on MCC metric.
Figure 17
Figure 17
CD diagram of comparing different prediction models with multivariate input pairwisely over the data contributors of Ohio_2018 dataset for the 60-min prediction horizon based on SE metric.
Figure 18
Figure 18
CD diagram of comparing different prediction models with multivariate input pairwisely over the data contributors of Ohio_2020 dataset for the 60-min prediction horizon based on RMSE metric.
Figure 19
Figure 19
CD diagram of comparing different prediction models with multivariate input pairwisely over the data contributors of Ohio_2020 dataset for the 60-min prediction horizon based on MAE metric.
Figure 20
Figure 20
CD diagram of comparing different prediction models with multivariate input pairwisely over the data contributors of Ohio_2020 dataset for the 60-min prediction horizon based on MCC metric.
Figure 21
Figure 21
CD diagram of comparing different prediction models with multivariate input pairwisely over the data contributors of Ohio_2020 dataset for the 60-min prediction horizon based on SE metric.
See this image and copyright information in PMC

References

    1. Kavakiotis, I. et al. Machine learning and data mining methods in diabetes research. Comput. Struct. Biotechnol. J.15, 104–116 (2017). - PMC - PubMed
    1. Woldaregay, A. Z. et al. Data-driven blood glucose pattern classification and anomalies detection: Machine-learning applications in type 1 diabetes. J. Med. Internet Res.21(5), 11030 (2019). - PMC - PubMed
    1. Cobelli, C., Renard, E. & Kovatchev, B. Artificial pancreas: Past, present, future. Diabetes60(11), 2672–2682 (2011). - PMC - PubMed
    1. Lunze, K., Singh, T., Walter, M., Brendel, M. D. & Leonhardt, S. Blood glucose control algorithms for type 1 diabetic patients: A methodological review. Biomed. Signal Process. Control8(2), 107–119 (2013).
    1. Woldaregay, A. Z. et al. Data-driven modeling and prediction of blood glucose dynamics: Machine learning applications in type 1 diabetes. Artif. Intell. Med.98, 109–134 (2019). - PubMed

Publication types

LinkOut - more resources