Heterogeneous temporal representation for diabetic blood glucose prediction
Background and aims: Blood glucose prediction (BGP) has increasingly been adopted for personalized monitoring of blood glucose levels in diabetic patients, providing valuable support for physicians in diagnosis and treatment planning. Despite the remarkable success achieved, applying BGP in multi-pa...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2023-07-01
|
| Series: | Frontiers in Physiology |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fphys.2023.1225638/full |
| _version_ | 1827899031275175936 |
|---|---|
| author | Yaohui Huang Yaohui Huang Zhikai Ni Zhenkun Lu Zhenkun Lu Xinqi He Jinbo Hu Boxuan Li Houguan Ya Yunxian Shi |
| author_facet | Yaohui Huang Yaohui Huang Zhikai Ni Zhenkun Lu Zhenkun Lu Xinqi He Jinbo Hu Boxuan Li Houguan Ya Yunxian Shi |
| author_sort | Yaohui Huang |
| collection | DOAJ |
| description | Background and aims: Blood glucose prediction (BGP) has increasingly been adopted for personalized monitoring of blood glucose levels in diabetic patients, providing valuable support for physicians in diagnosis and treatment planning. Despite the remarkable success achieved, applying BGP in multi-patient scenarios remains problematic, largely due to the inherent heterogeneity and uncertain nature of continuous glucose monitoring (CGM) data obtained from diverse patient profiles.Methodology: This study proposes the first graph-based Heterogeneous Temporal Representation (HETER) network for multi-patient Blood Glucose Prediction (BGP). Specifically, HETER employs a flexible subsequence repetition method (SSR) to align the heterogeneous input samples, in contrast to the traditional padding or truncation methods. Then, the relationships between multiple samples are constructed as a graph and learned by HETER to capture global temporal characteristics. Moreover, to address the limitations of conventional graph neural networks in capturing local temporal dependencies and providing linear representations, HETER incorporates both a temporally-enhanced mechanism and a linear residual fusion into its architecture.Results: Comprehensive experiments were conducted to validate the proposed method using real-world data from 112 patients in two hospitals, comparing it with five well-known baseline methods. The experimental results verify the robustness and accuracy of the proposed HETER, which achieves the maximal improvement of 31.42%, 27.18%, and 34.85% in terms of MAE, MAPE, and RMSE, respectively, over the second-best comparable method.Discussions: HETER integrates global and local temporal information from multi-patient samples to alleviate the impact of heterogeneity and uncertainty. This method can also be extended to other clinical tasks, thereby facilitating efficient and accurate capture of crucial pattern information in structured medical data. |
| first_indexed | 2024-03-12T23:11:28Z |
| format | Article |
| id | doaj.art-54c582c99569403694b28a5a3287c4e5 |
| institution | Directory Open Access Journal |
| issn | 1664-042X |
| language | English |
| last_indexed | 2024-03-12T23:11:28Z |
| publishDate | 2023-07-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Physiology |
| spelling | doaj.art-54c582c99569403694b28a5a3287c4e52023-07-18T01:03:35ZengFrontiers Media S.A.Frontiers in Physiology1664-042X2023-07-011410.3389/fphys.2023.12256381225638Heterogeneous temporal representation for diabetic blood glucose predictionYaohui Huang0Yaohui Huang1Zhikai Ni2Zhenkun Lu3Zhenkun Lu4Xinqi He5Jinbo Hu6Boxuan Li7Houguan Ya8Yunxian Shi9College of Electronic Information, Guangxi Minzu University, Nanning, ChinaLaboratory of Intelligent Information Processing and Intelligent Medical, Guangxi Minzu University, Nanning, ChinaDepartment of Electronic Science, Xiamen University, Xiamen, ChinaCollege of Electronic Information, Guangxi Minzu University, Nanning, ChinaLaboratory of Intelligent Information Processing and Intelligent Medical, Guangxi Minzu University, Nanning, ChinaCollege of Electronic Information, Guangxi Minzu University, Nanning, ChinaCollege of Electronic Information, Guangxi Minzu University, Nanning, ChinaCollege of Electronic Information, Guangxi Minzu University, Nanning, ChinaCollege of Electronic Information, Guangxi Minzu University, Nanning, ChinaCollege of Electronic Information, Guangxi Minzu University, Nanning, ChinaBackground and aims: Blood glucose prediction (BGP) has increasingly been adopted for personalized monitoring of blood glucose levels in diabetic patients, providing valuable support for physicians in diagnosis and treatment planning. Despite the remarkable success achieved, applying BGP in multi-patient scenarios remains problematic, largely due to the inherent heterogeneity and uncertain nature of continuous glucose monitoring (CGM) data obtained from diverse patient profiles.Methodology: This study proposes the first graph-based Heterogeneous Temporal Representation (HETER) network for multi-patient Blood Glucose Prediction (BGP). Specifically, HETER employs a flexible subsequence repetition method (SSR) to align the heterogeneous input samples, in contrast to the traditional padding or truncation methods. Then, the relationships between multiple samples are constructed as a graph and learned by HETER to capture global temporal characteristics. Moreover, to address the limitations of conventional graph neural networks in capturing local temporal dependencies and providing linear representations, HETER incorporates both a temporally-enhanced mechanism and a linear residual fusion into its architecture.Results: Comprehensive experiments were conducted to validate the proposed method using real-world data from 112 patients in two hospitals, comparing it with five well-known baseline methods. The experimental results verify the robustness and accuracy of the proposed HETER, which achieves the maximal improvement of 31.42%, 27.18%, and 34.85% in terms of MAE, MAPE, and RMSE, respectively, over the second-best comparable method.Discussions: HETER integrates global and local temporal information from multi-patient samples to alleviate the impact of heterogeneity and uncertainty. This method can also be extended to other clinical tasks, thereby facilitating efficient and accurate capture of crucial pattern information in structured medical data.https://www.frontiersin.org/articles/10.3389/fphys.2023.1225638/fullcontinuous glucose monitoringdiabetes mellitustime seriespredictiongraph neural networkdeep neural network |
| spellingShingle | Yaohui Huang Yaohui Huang Zhikai Ni Zhenkun Lu Zhenkun Lu Xinqi He Jinbo Hu Boxuan Li Houguan Ya Yunxian Shi Heterogeneous temporal representation for diabetic blood glucose prediction Frontiers in Physiology continuous glucose monitoring diabetes mellitus time series prediction graph neural network deep neural network |
| title | Heterogeneous temporal representation for diabetic blood glucose prediction |
| title_full | Heterogeneous temporal representation for diabetic blood glucose prediction |
| title_fullStr | Heterogeneous temporal representation for diabetic blood glucose prediction |
| title_full_unstemmed | Heterogeneous temporal representation for diabetic blood glucose prediction |
| title_short | Heterogeneous temporal representation for diabetic blood glucose prediction |
| title_sort | heterogeneous temporal representation for diabetic blood glucose prediction |
| topic | continuous glucose monitoring diabetes mellitus time series prediction graph neural network deep neural network |
| url | https://www.frontiersin.org/articles/10.3389/fphys.2023.1225638/full |
| work_keys_str_mv | AT yaohuihuang heterogeneoustemporalrepresentationfordiabeticbloodglucoseprediction AT yaohuihuang heterogeneoustemporalrepresentationfordiabeticbloodglucoseprediction AT zhikaini heterogeneoustemporalrepresentationfordiabeticbloodglucoseprediction AT zhenkunlu heterogeneoustemporalrepresentationfordiabeticbloodglucoseprediction AT zhenkunlu heterogeneoustemporalrepresentationfordiabeticbloodglucoseprediction AT xinqihe heterogeneoustemporalrepresentationfordiabeticbloodglucoseprediction AT jinbohu heterogeneoustemporalrepresentationfordiabeticbloodglucoseprediction AT boxuanli heterogeneoustemporalrepresentationfordiabeticbloodglucoseprediction AT houguanya heterogeneoustemporalrepresentationfordiabeticbloodglucoseprediction AT yunxianshi heterogeneoustemporalrepresentationfordiabeticbloodglucoseprediction |