Exploring Feasibility of Multivariate Deep Learning Models in Predicting COVID-19 Epidemic
Background: Mathematical models are powerful tools to study COVID-19. However, one fundamental challenge in current modeling approaches is the lack of accurate and comprehensive data. Complex epidemiological systems such as COVID-19 are especially challenging to the commonly used mechanistic model w...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2021-07-01
|
| Series: | Frontiers in Public Health |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fpubh.2021.661615/full |
| _version_ | 1818589425095933952 |
|---|---|
| author | Shi Chen Shi Chen Rajib Paul Rajib Paul Daniel Janies Keith Murphy Tinghao Feng Jean-Claude Thill Jean-Claude Thill |
| author_facet | Shi Chen Shi Chen Rajib Paul Rajib Paul Daniel Janies Keith Murphy Tinghao Feng Jean-Claude Thill Jean-Claude Thill |
| author_sort | Shi Chen |
| collection | DOAJ |
| description | Background: Mathematical models are powerful tools to study COVID-19. However, one fundamental challenge in current modeling approaches is the lack of accurate and comprehensive data. Complex epidemiological systems such as COVID-19 are especially challenging to the commonly used mechanistic model when our understanding of this pandemic rapidly refreshes.Objective: We aim to develop a data-driven workflow to extract, process, and develop deep learning (DL) methods to model the COVID-19 epidemic. We provide an alternative modeling approach to complement the current mechanistic modeling paradigm.Method: We extensively searched, extracted, and annotated relevant datasets from over 60 official press releases in Hubei, China, in 2020. Multivariate long short-term memory (LSTM) models were developed with different architectures to track and predict multivariate COVID-19 time series for 1, 2, and 3 days ahead. As a comparison, univariate LSTMs were also developed to track new cases, total cases, and new deaths.Results: A comprehensive dataset with 10 variables was retrieved and processed for 125 days in Hubei. Multivariate LSTM had reasonably good predictability on new deaths, hospitalization of both severe and critical patients, total discharges, and total monitored in hospital. Multivariate LSTM showed better results for new and total cases, and new deaths for 1-day-ahead prediction than univariate counterparts, but not for 2-day and 3-day-ahead predictions. Besides, more complex LSTM architecture seemed not to increase overall predictability in this study.Conclusion: This study demonstrates the feasibility of DL models to complement current mechanistic approaches when the exact epidemiological mechanisms are still under investigation. |
| first_indexed | 2024-12-16T09:40:26Z |
| format | Article |
| id | doaj.art-3122be121df44231810f12b145f73004 |
| institution | Directory Open Access Journal |
| issn | 2296-2565 |
| language | English |
| last_indexed | 2024-12-16T09:40:26Z |
| publishDate | 2021-07-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Public Health |
| spelling | doaj.art-3122be121df44231810f12b145f730042022-12-21T22:36:17ZengFrontiers Media S.A.Frontiers in Public Health2296-25652021-07-01910.3389/fpubh.2021.661615661615Exploring Feasibility of Multivariate Deep Learning Models in Predicting COVID-19 EpidemicShi Chen0Shi Chen1Rajib Paul2Rajib Paul3Daniel Janies4Keith Murphy5Tinghao Feng6Jean-Claude Thill7Jean-Claude Thill8Department of Public Health Sciences, University of North Carolina at Charlotte, Charlotte, NC, United StatesSchool of Data Science, University of North Carolina at Charlotte, Charlotte, NC, United StatesDepartment of Public Health Sciences, University of North Carolina at Charlotte, Charlotte, NC, United StatesSchool of Data Science, University of North Carolina at Charlotte, Charlotte, NC, United StatesDepartment of Bioinformatics and Genomics, University of North Carolina at Charlotte, Charlotte, NC, United StatesDepartment of Public Health Sciences, University of North Carolina at Charlotte, Charlotte, NC, United StatesDepartment of Computer Science, University of North Carolina at Charlotte, Charlotte, NC, United StatesSchool of Data Science, University of North Carolina at Charlotte, Charlotte, NC, United StatesDepartment of Geography and Earth Sciences, University of North Carolina at Charlotte, Charlotte, NC, United StatesBackground: Mathematical models are powerful tools to study COVID-19. However, one fundamental challenge in current modeling approaches is the lack of accurate and comprehensive data. Complex epidemiological systems such as COVID-19 are especially challenging to the commonly used mechanistic model when our understanding of this pandemic rapidly refreshes.Objective: We aim to develop a data-driven workflow to extract, process, and develop deep learning (DL) methods to model the COVID-19 epidemic. We provide an alternative modeling approach to complement the current mechanistic modeling paradigm.Method: We extensively searched, extracted, and annotated relevant datasets from over 60 official press releases in Hubei, China, in 2020. Multivariate long short-term memory (LSTM) models were developed with different architectures to track and predict multivariate COVID-19 time series for 1, 2, and 3 days ahead. As a comparison, univariate LSTMs were also developed to track new cases, total cases, and new deaths.Results: A comprehensive dataset with 10 variables was retrieved and processed for 125 days in Hubei. Multivariate LSTM had reasonably good predictability on new deaths, hospitalization of both severe and critical patients, total discharges, and total monitored in hospital. Multivariate LSTM showed better results for new and total cases, and new deaths for 1-day-ahead prediction than univariate counterparts, but not for 2-day and 3-day-ahead predictions. Besides, more complex LSTM architecture seemed not to increase overall predictability in this study.Conclusion: This study demonstrates the feasibility of DL models to complement current mechanistic approaches when the exact epidemiological mechanisms are still under investigation.https://www.frontiersin.org/articles/10.3389/fpubh.2021.661615/fullCOVID-19epidemicmodelingdeep learningmultivariate |
| spellingShingle | Shi Chen Shi Chen Rajib Paul Rajib Paul Daniel Janies Keith Murphy Tinghao Feng Jean-Claude Thill Jean-Claude Thill Exploring Feasibility of Multivariate Deep Learning Models in Predicting COVID-19 Epidemic Frontiers in Public Health COVID-19 epidemic modeling deep learning multivariate |
| title | Exploring Feasibility of Multivariate Deep Learning Models in Predicting COVID-19 Epidemic |
| title_full | Exploring Feasibility of Multivariate Deep Learning Models in Predicting COVID-19 Epidemic |
| title_fullStr | Exploring Feasibility of Multivariate Deep Learning Models in Predicting COVID-19 Epidemic |
| title_full_unstemmed | Exploring Feasibility of Multivariate Deep Learning Models in Predicting COVID-19 Epidemic |
| title_short | Exploring Feasibility of Multivariate Deep Learning Models in Predicting COVID-19 Epidemic |
| title_sort | exploring feasibility of multivariate deep learning models in predicting covid 19 epidemic |
| topic | COVID-19 epidemic modeling deep learning multivariate |
| url | https://www.frontiersin.org/articles/10.3389/fpubh.2021.661615/full |
| work_keys_str_mv | AT shichen exploringfeasibilityofmultivariatedeeplearningmodelsinpredictingcovid19epidemic AT shichen exploringfeasibilityofmultivariatedeeplearningmodelsinpredictingcovid19epidemic AT rajibpaul exploringfeasibilityofmultivariatedeeplearningmodelsinpredictingcovid19epidemic AT rajibpaul exploringfeasibilityofmultivariatedeeplearningmodelsinpredictingcovid19epidemic AT danieljanies exploringfeasibilityofmultivariatedeeplearningmodelsinpredictingcovid19epidemic AT keithmurphy exploringfeasibilityofmultivariatedeeplearningmodelsinpredictingcovid19epidemic AT tinghaofeng exploringfeasibilityofmultivariatedeeplearningmodelsinpredictingcovid19epidemic AT jeanclaudethill exploringfeasibilityofmultivariatedeeplearningmodelsinpredictingcovid19epidemic AT jeanclaudethill exploringfeasibilityofmultivariatedeeplearningmodelsinpredictingcovid19epidemic |