Blood Glucose Concentration Prediction Based on Double Decomposition and Deep Extreme Learning Machine Optimized by Nonlinear Marine Predator Algorithm
Abstract
:1. Introduction
- (a)
- In the forecasting of short-term blood glucose concentration data, the double decomposition method is adopted to decompose the original blood glucose sequence and the generated residual component, effectively reducing the non-stationarity and non-linearity of the sequence. The overall prediction accuracy of the model with full consideration of residual component is higher.
- (b)
- The DELM network is utilized for predicting blood glucose data, which significantly improves the ability to acquire the characteristics of blood glucose concentration in depth. Furthermore, the training procedure of DELM is relatively uncomplicated, and the random initialization method accelerates the training speed, so as to achieve the high-precision prediction effect in a short time.
- (c)
- The NMPA algorithm adopts new parameters to adjust the predator’s moving step size in the process of optimization, which can maintain a high exploration ability. In the second stage, a control parameter is introduced to balance the development and exploration aspects of the algorithm, and further enhance the global and local search capabilities. Compared with the MPA algorithm, the probability of realizing that the global optimal value is greatly improved.
- (d)
- The NMPA algorithm is employed to optimize the weight parameters of DELM network and establish the optimal parameter combination model, which is helpful to avoid large fluctuations in the forecasting outcomes, improve the predictive performance of the model and make it more stable and effective.
2. Decomposition and Prediction Algorithm Theory
2.1. VMD Algorithm
2.2. TVF-EMD Algorithm
- (1)
- Compute the local cutoff frequency.
- (2)
- Signal reconstruction.
- (3)
- Solve the cut-off criterion
2.3. Deep Extreme Learning Machine
2.4. Marine Predator Algorithm
3. Construction of the Model Using Double Decomposition and Deep Extreme Learning Machine Optimized by Nonlinear Marine Predator Algorithm
3.1. Double Decomposition Strategy
3.2. NMPA-DELM Method
3.2.1. Nonlinear Marine Predator Algorithm
3.2.2. Construction of NMPA-DELM
- (1)
- Set the parameters of NMPA, encompassing the population number , maximum number of iterations , variable dimension and search scope .
- (2)
- Population initialization: The ownership values are combined as the position of each prey, and the position of the individual prey population is initialized by Formula (14).
- (3)
- The mean square error between the real value and the predicted value is used as the fitness function to calculate and rank the fitness value of the prey individual. Through this process, the individual with the best fitness is found out and the elite matrix is constructed.
- (4)
- The prey position is updated in the high-speed ratio stage, equal-speed ratio stage and low-speed ratio stage, memory storage and elite update are realized and new adaptive moving step parameters and inertia weights are calculated. For vortex or FAD effects, the prey is updated using Equation (34).
- (5)
- Determine whether the maximum number of iterations has been reached. If yes, the iteration ends and the best elite matrix is entered into the DELM as a combination of weights. Otherwise, skip to step 3.
3.3. Prediction of Blood Glucose Concentration Based on Double Decomposition and NMPA-DELM
- (1)
- Decompose the original blood sugar concentration data using the VMD method to acquire a series of smoother modal components.
- (2)
- The residual component after VMD processing is calculated and the component is decomposed using the TVF-EMD method. In this way, the complexity of the fluctuation of the residual component is reduced and the double decomposition results are obtained.
- (3)
- The NMPA algorithm is employed to optimize the weight parameters of the DELM network and find the most effective parameter combination, so as to avoid the volatility of the network prediction performance.
- (4)
- Input all the components into the NMPA-DELM network for training separately to obtain the predicted values of each subsequence.
- (5)
- Merge the forecasting outcomes of all sub-sequences to obtain the ultimate predicted sequence of blood sugar concentration.
4. Experimental Simulation Analysis
4.1. Sources and Configurations for Blood Glucose Concentration Data
4.2. Model Performance Evaluation Index
4.3. Ablation Study
4.4. Comparison Between NMPA Algorithm and Other Optimization Algorithms
4.5. Comparison Between VMD-TVF-EMD-NMPA-DELM Model Versus Other Hybrid Models
4.6. Statistical Robustness Analysis and Diebold–Mariano Test
4.7. Real-Time Analysis
5. Conclusions
Author Contributions
Funding
Data Availability Statement
Conflicts of Interest
References
- Arora, S.; Kumar, S.; Kumar, P. Multivariate models of blood glucose prediction in type1 diabetes: A survey of the state-of-the-art. Curr. Pharm. Biotechnol. 2023, 24, 532–552. [Google Scholar] [PubMed]
- Wang, Y.; Wang, T. Application of improved LightGBM model in blood glucose prediction. Appl. Sci. 2020, 10, 3227. [Google Scholar] [CrossRef]
- Prendin, F.; Pavan, J.; Cappon, G.; Del Favero, S.; Sparacino, G.; Facchinetti, A. The importance of interpreting machine learning models for blood glucose prediction in diabetes: An analysis using SHAP. Sci. Rep. 2023, 13, 16865. [Google Scholar] [CrossRef] [PubMed]
- Zhu, T.; Wang, W.; Yu, M. A novel blood glucose time series prediction framework based on a novel signal decomposition method. Chaos Solitons Fractals 2022, 164, 112673. [Google Scholar] [CrossRef]
- Lu, H.Y.; Ding, X.; Hirst, J.E.; Yang, Y.; Yang, J.; Mackillop, L.; Clifton, D.A. Digital health and machine learning technologies for blood glucose monitoring and management of gestational diabetes. IEEE Rev. Biomed. Eng. 2023, 17, 98–117. [Google Scholar] [CrossRef]
- Xie, J.; Wang, Q. Benchmarking machine learning algorithms on blood glucose prediction for type I diabetes in comparison with classical time-series models. IEEE Trans. Biomed. Eng. 2020, 67, 3101–3124. [Google Scholar] [CrossRef]
- Sevil, M.; Rashid, M.; Hajizadeh, I.; Park, M.; Quinn, L.; Cinar, A. Physical activity and psychological stress detection and assessment of their effects on glucose concentration predictions in diabetes management. IEEE Trans. Biomed. Eng. 2021, 68, 2251–2260. [Google Scholar] [CrossRef]
- Rabby, M.F.; Tu, Y.; Hossen, M.I.; Lee, I.; Maida, A.S.; Hei, X. Stacked LSTM based deep recurrent neural network with kalman smoothing for blood glucose prediction. BMC Med. Inform. Decis. Mak. 2021, 21, 101. [Google Scholar] [CrossRef]
- Alfian, G.; Syafrudin, M.; Anshari, M.; Benes, F.; Atmaji, F.T.D.; Fahrurrozi, I.; Hidayatullah, A.F.; Rhee, J. Blood glucose prediction model for type 1 diabetes based on artificial neural network with time-domain features. Biocybern. Biomed. Eng. 2020, 40, 1586–1599. [Google Scholar] [CrossRef]
- Li, N.; Tuo, J.; Wang, Y.; Wang, M. Prediction of blood glucose concentration for type 1 diabetes based on echo state networks embedded with incremental learning. Neurocomputing 2020, 378, 248–259. [Google Scholar] [CrossRef]
- Iacono, F.; Magni, L.; Toffanin, C. Personalized LSTM-based alarm systems for hypoglycemia and hyperglycemia prevention. Biomed. Signal Process. Control 2023, 86, 105167. [Google Scholar] [CrossRef]
- Jaloli, M.; Cescon, M. Long-term prediction of blood glucose levels in type 1 diabetes using a CNN-LSTM-based deep neural network. J. Diabetes Sci. Technol. 2023, 17, 1590–1601. [Google Scholar] [CrossRef] [PubMed]
- Dudukcu, H.V.; Taskiran, M.; Yildirim, T. Blood glucose prediction with deep neural networks using weighted decision level fusion. Biocybern. Biomed. Eng. 2021, 41, 1208–1223. [Google Scholar] [CrossRef]
- Saravanan, R.; Mahmud, F. Blood Glucose Prediction Based on ARIMA Time-Series Machine Learning Model. Evol. Electr. Electron. Eng. 2023, 4, 457–463. [Google Scholar]
- Ding, S.; Zhao, H.; Zhang, Y.; Xu, X.; Nie, R. Extreme learning machine: Algorithm, theory and applications. Artif. Intell. Rev. 2015, 44, 103–115. [Google Scholar] [CrossRef]
- Daniels, J.; Herrero, P.; Georgiou, P. A multitask learning approach to personalized blood glucose prediction. IEEE J. Biomed. Health Inform. 2021, 26, 436–445. [Google Scholar] [CrossRef]
- Martinsson, J.; Schliep, A.; Eliasson, B.; Mogren, O. Blood glucose prediction with variance estimation using recurrent neural networks. J. Healthc. Inform. Res. 2020, 4, 1–18. [Google Scholar] [CrossRef]
- Wang, W.; Tong, M.; Yu, M. Blood glucose prediction with VMD and LSTM optimized by improved particle swarm optimization. IEEE Access 2020, 8, 217908–217916. [Google Scholar] [CrossRef]
- Yang, T.; Yu, X.; Ma, N.; Wu, R.; Li, H. An autonomous channel deep learning framework for blood glucose prediction. Appl. Soft Comput. 2022, 120, 108636. [Google Scholar] [CrossRef]
- Zhu, T.; Li, K.; Chen, J.; Herrero, P.; Georgiou, P. Dilated recurrent neural networks for glucose forecasting in type 1 diabetes. J. Healthc. Inform. Res. 2020, 4, 308–324. [Google Scholar] [CrossRef]
- Yang, G.; Liu, S.; Li, Y.; He, L. Short-term prediction method of blood glucose based on temporal multi-head attention mechanism for diabetic patients. Biomed. Signal Process. Control 2023, 82, 104552. [Google Scholar] [CrossRef]
- Kumari, G.L.A.; Padmaja, P.; Suma, J.G. A novel method for prediction of diabetes mellitus using deep convolutional neural network and long short-term memory. Indones. J. Electr. Eng. Comput. Sci. 2022, 26, 404–413. [Google Scholar] [CrossRef]
- Long, X.; Wang, J.; Gong, S.; Li, G.; Ju, H. Reference evapotranspiration estimation using long short-term memory network and wavelet-coupled long short-term memory network. Irrig. Drain. 2022, 71, 855–881. [Google Scholar] [CrossRef]
- Meng, A.; Zhu, Z.; Deng, W.; Ou, Z.; Lin, S.; Wang, C.; Xu, X.; Wang, X.; Yin, H.; Luo, J. A novel wind power prediction approach using multivariate variational mode decomposition and multi-objective crisscross optimization based deep extreme learning machine. Energy 2022, 260, 124957. [Google Scholar] [CrossRef]
- Siddiqui, S.Y.; Khan, M.A.; Abbas, S.; Khan, F. Smart occupancy detection for road traffic parking using deep extreme learning machine. J. King Saud Univ. Comput. Inf. Sci. 2022, 34, 727–733. [Google Scholar]
- Xiong, J.; Liang, W.; Liang, X.; Yao, J. Intelligent quantification of natural gas pipeline defects using improved sparrow search algorithm and deep extreme learning machine. Chem. Eng. Res. Des. 2022, 183, 567–579. [Google Scholar] [CrossRef]
- Song, C.; Yao, L. Application of artificial intelligence based on synchrosqueezed wavelet transform and improved deep extreme learning machine in water quality prediction. Environ. Sci. Pollut. Res. 2022, 29, 38066–38082. [Google Scholar] [CrossRef]
- Sadiq, A.S.; Dehkordi, A.A.; Mirjalili, S.; Pham, Q.V. Nonlinear marine predator algorithm: A cost-effective optimizer for fair power allocation in NOMA-VLC-B5G networks. Expert Syst. Appl. 2022, 203, 117395. [Google Scholar] [CrossRef]
- Garai, S.; Paul, R.K.; Yeasin, M.; Paul, A.K. CEEMDAN-Based Hybrid Machine Learning Models for Time Series Forecasting Using MARS Algorithm and PSO-Optimization. Neural Process. Lett. 2024, 56, 92. [Google Scholar] [CrossRef]
- Jian, X.; Xia, Y.; Sun, L. An indirect method for bridge mode shapes identification based on wavelet analysis. Struct. Control Health Monit. 2020, 27, e2630. [Google Scholar] [CrossRef]
- Zhang, Y.; Li, C.; Jiang, Y.; Sun, L.; Zhao, R.; Yan, K.; Wang, W. Accurate prediction of water quality in urban drainage network with integrated EMD-LSTM model. J. Clean. Prod. 2022, 354, 131724. [Google Scholar] [CrossRef]
- Zuo, G.; Luo, J.; Wang, N.; Lian, Y.; He, X. Decomposition ensemble model based on variational mode decomposition and long short-term memory for streamflow forecasting. J. Hydrol. 2020, 585, 124776. [Google Scholar] [CrossRef]
- Gu, J.; Peng, Y.; Lu, H.; Chang, X.; Cao, S.; Chen, G.; Cao, B. An optimized variational mode decomposition method and its application in vibration signal analysis of bearings. Struct. Health Monit. 2022, 21, 2386–2407. [Google Scholar] [CrossRef]
- Gan, M.; Pan, H.; Chen, Y.; Pan, S. Application of the Variational Mode Decomposition (VMD) method to river tides. Estuar. Coast. Shelf Sci. 2021, 261, 107570. [Google Scholar] [CrossRef]
- Li, H.; Li, Z.; Mo, W. A time varying filter approach for empirical mode decomposition. Signal Process. 2017, 138, 146–158. [Google Scholar] [CrossRef]
- Rehman, A.; Athar, A.; Khan, M.A.; Abbas, S.; Fatima, A.; Saeed, A. Modelling, simulation, and optimization of diabetes type II prediction using deep extreme learning machine. J. Ambient Intell. Smart Environ. 2020, 12, 125–138. [Google Scholar] [CrossRef]
- Faramarzi, A.; Heidarinejad, M.; Mirjalili, S.; Gandomi, A.H. Marine Predators Algorithm: A nature-inspired metaheuristic. Expert Syst. Appl. 2020, 152, 113377. [Google Scholar] [CrossRef]
- Wang, W.; Shen, Y.; Chen, G. Blood glucose concentration prediction based on VMD-KELM-AdaBoost. Med. Biol. Eng. Comput. 2021, 59, 2219–2235. [Google Scholar]
- Ding, L.; Bai, Y.; Liu, M.D.; Fan, M.H.; Yang, J. Predicting short wind speed with a hybrid model based on a piecewise error correction method and Elman neural network. Energy 2022, 244, 122630. [Google Scholar] [CrossRef]
- Noorollahi, Y.; Jokar, M.A.; Kalhor, A. Using artificial neural networks for temporal and spatial wind speed forecasting in Iran. Energy Convers. Manag. 2016, 115, 17–25. [Google Scholar] [CrossRef]
Patient No. | #12 | #71 | #84 | ||||||
---|---|---|---|---|---|---|---|---|---|
Evaluation index | MAE (mg/dL) | RMSE (mg/dL) | MAPE (%) | MAE (mg/dL) | RMSE (mg/dL) | MAPE (%) | MAE (mg/dL) | RMSE (mg/dL) | MAPE (%) |
DELM | 19.0183 | 25.1349 | 12.5746 | 17.1421 | 24.8637 | 15.0656 | 25.9199 | 34.7961 | 18.1434 |
NMPA-DELM | 14.3498 | 22.3565 | 9.0162 | 15.6890 | 22.8202 | 13.3681 | 22.6633 | 31.4187 | 14.1642 |
VMD-NMPA-DELM | 4.8456 | 6.6736 | 3.1131 | 3.7409 | 5.3116 | 3.2536 | 5.6268 | 7.5108 | 3.7507 |
Proposed | 3.5205 | 5.2095 | 2.2836 | 2.7935 | 4.2410 | 2.3889 | 4.4200 | 6.3246 | 2.9702 |
Patient No. | #12 | #71 | #84 | ||||||
---|---|---|---|---|---|---|---|---|---|
Evaluation index | MAE (mg/dL) | RMSE (mg/dL) | MAPE (%) | MAE (mg/dL) | RMSE (mg/dL) | MAPE (%) | MAE (mg/dL) | RMSE (mg/dL) | MAPE (%) |
GA-DELM | 17.5289 | 23.9132 | 11.5425 | 16.6339 | 23.9651 | 14.6518 | 24.3874 | 33.1517 | 15.9302 |
PSO-DELM | 16.8212 | 23.3201 | 11.1404 | 16.5482 | 23.6498 | 14.3487 | 23.7039 | 32.7162 | 15.4212 |
SSA-DELM | 16.6695 | 23.2439 | 10.9725 | 16.4413 | 23.4639 | 14.2955 | 23.3347 | 32.4914 | 14.9953 |
MPA-DELM | 14.5976 | 22.8046 | 9.1169 | 16.0365 | 23.1997 | 13.8470 | 22.8014 | 31.9200 | 14.4662 |
NMPA-DELM | 14.3498 | 22.3565 | 9.0162 | 15.6890 | 22.8202 | 13.3681 | 22.6633 | 31.4187 | 14.1642 |
Patient No. | #12 | #71 | #84 | ||||||
---|---|---|---|---|---|---|---|---|---|
Evaluation index | MAE (mg/dL) | RMSE (mg/dL) | MAPE (%) | MAE (mg/dL) | RMSE (mg/dL) | MAPE (%) | MAE (mg/dL) | RMSE (mg/dL) | MAPE (%) |
Model 1 | 4.4791 | 6.2173 | 2.8718 | 3.4247 | 4.9861 | 2.9413 | 5.2656 | 7.0412 | 3.4811 |
Model 2 | 4.2762 | 5.9527 | 2.7493 | 3.3403 | 4.8325 | 2.8338 | 5.0098 | 6.7994 | 3.3634 |
Model 3 | 4.1585 | 5.7932 | 2.6799 | 3.2813 | 4.7174 | 2.7979 | 4.9443 | 6.7308 | 3.3274 |
Model 4 | 3.8372 | 5.4144 | 2.4815 | 3.0843 | 4.4534 | 2.6252 | 4.7839 | 6.5851 | 3.2077 |
Proposed | 3.5205 | 5.2095 | 2.2836 | 2.7935 | 4.2410 | 2.3889 | 4.4200 | 6.3246 | 2.9702 |
Model | MAE (mg/dL) | RMSE (mg/dL) | MAPE (%) |
---|---|---|---|
DELM | 23.3270 | 30.9217 | 15.5808 |
GA-DELM | 21.3548 | 28.7121 | 13.8196 |
PSO-DELM | 20.4191 | 28.1099 | 13.2968 |
SSA-DELM | 19.8517 | 27.8276 | 13.0386 |
MPA-DELM | 19.7947 | 27.6643 | 12.8718 |
NMPA-DELM | 18.9199 | 27.0617 | 12.4387 |
VMD-NMPA-DELM | 5.6831 | 7.6043 | 3.7989 |
VMD-TVF-EMD-NMPA-ELM | 4.9396 | 6.7904 | 3.2331 |
POA-TVF-EMD-CNN-LSTM-EC | 4.6880 | 6.4030 | 3.0907 |
CEEMDAN-SE-EC-BiLSTM | 4.6280 | 6.3193 | 3.0553 |
VMD-TVF-EMD-NMPA-LSTM | 4.4545 | 6.0826 | 2.9268 |
VMD-TVF-EMD-NMPA-DELM | 4.2239 | 5.8960 | 2.7805 |
Model | DM Test | p-Value |
---|---|---|
DELM | 10.3235 | 1.662 × 10−25 |
GA-DELM | 8.7042 | 1.013 × 10−18 |
PSO-DELM | 8.3746 | 4.385 × 10−17 |
SSA-DELM | 8.2487 | 3.481 × 10−16 |
MPA-DELM | 7.1218 | 1.179 × 10−12 |
NMPA-DELM | 7.0338 | 2.242 × 10−12 |
VMD-NMPA-DELM | 6.8733 | 1.116 × 10−11 |
VMD-TVF-EMD-NMPA-ELM | 5.9251 | 3.053 × 10−9 |
POA-TVF-EMD-CNN-LSTM-EC | 5.3039 | 1.133 × 10−7 |
CEEMDAN-SE-EC-BiLSTM | 4.6235 | 3.773 × 10−6 |
VMD-TVF-EMD-NMPA-LSTM | 3.5113 | 4.459 × 10−4 |
Model | Training Stage t/s | Execution Stage t/s |
---|---|---|
DELM | 1.9537 | 0.2062 |
GA-DELM | 22.8169 | 0.2986 |
PSO-DELM | 20.5278 | 0.2765 |
SSA-DELM | 21.4839 | 0.2903 |
MPA-DELM | 18.0356 | 0.2607 |
NMPA-DELM | 17.5269 | 0.2425 |
VMD-NMPA-DELM | 56.9832 | 0.5483 |
VMD-TVF-EMD-NMPA-ELM | 95.3467 | 0.9245 |
POA-TVF-EMD-CNN-LSTM-EC | 289.1745 | 2.3078 |
CEEMDAN-SE-EC-BiLSTM | 312.6472 | 2.5391 |
VMD-TVF-EMD-NMPA-LSTM | 357.8631 | 2.8146 |
VMD-TVF-EMD-NMPA-DELM | 101.4359 | 1.0821 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Shen, Y.; Li, D.; Wang, W.; Dong, X. Blood Glucose Concentration Prediction Based on Double Decomposition and Deep Extreme Learning Machine Optimized by Nonlinear Marine Predator Algorithm. Mathematics 2024, 12, 3708. https://doi.org/10.3390/math12233708
Shen Y, Li D, Wang W, Dong X. Blood Glucose Concentration Prediction Based on Double Decomposition and Deep Extreme Learning Machine Optimized by Nonlinear Marine Predator Algorithm. Mathematics. 2024; 12(23):3708. https://doi.org/10.3390/math12233708
Chicago/Turabian StyleShen, Yang, Deyi Li, Wenbo Wang, and Xu Dong. 2024. "Blood Glucose Concentration Prediction Based on Double Decomposition and Deep Extreme Learning Machine Optimized by Nonlinear Marine Predator Algorithm" Mathematics 12, no. 23: 3708. https://doi.org/10.3390/math12233708
APA StyleShen, Y., Li, D., Wang, W., & Dong, X. (2024). Blood Glucose Concentration Prediction Based on Double Decomposition and Deep Extreme Learning Machine Optimized by Nonlinear Marine Predator Algorithm. Mathematics, 12(23), 3708. https://doi.org/10.3390/math12233708