Developing a multi-variate prediction model for COVID-19 from crowd-sourced respiratory voice data
Aim: COVID-19 has affected more than 223 countries worldwide and in the post-COVID era, there is a pressing need for non-invasive, low-cost, and highly scalable solutions to detect COVID-19. This study focuses on the analysis of voice features and machine learning models in the automatic detection o...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Open Exploration Publishing Inc.
2024-08-01
|
| Series: | Exploration of Digital Health Technologies |
| Subjects: | |
| Online Access: | https://www.explorationpub.com/uploads/Article/A101122/101122.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849471748318167040 |
|---|---|
| author | Yuyang Yan Wafaa Aljbawi Sami O. Simons Visara Urovi |
| author_facet | Yuyang Yan Wafaa Aljbawi Sami O. Simons Visara Urovi |
| author_sort | Yuyang Yan |
| collection | DOAJ |
| description | Aim: COVID-19 has affected more than 223 countries worldwide and in the post-COVID era, there is a pressing need for non-invasive, low-cost, and highly scalable solutions to detect COVID-19. This study focuses on the analysis of voice features and machine learning models in the automatic detection of COVID-19. Methods: We develop a deep learning model to identify COVID-19 from voice recording data. The novelty of this work is in the development of deep learning models for COVID-19 identification from only voice recordings. We use the Cambridge COVID-19 Sound database which contains 893 speech samples, crowd-sourced from 4,352 participants via a COVID-19 Sounds app. Voice features including Mel-spectrograms and Mel-frequency cepstral coefficients (MFCC) and convolutional neural network (CNN) Encoder features are extracted. Based on the voice data, we develop deep learning classification models to detect COVID-19 cases. These models include long short-term memory (LSTM), CNN and Hidden-Unit BERT (HuBERT). Results: We compare their predictive power to baseline machine learning models. HuBERT achieves the highest accuracy of 86% and the highest AUC of 0.93. Conclusions: The results achieved with the proposed models suggest promising results in COVID-19 diagnosis from voice recordings when compared to the results obtained from the state-of-the-art. |
| format | Article |
| id | doaj-art-41e8b358a71b4f8abbb07faf5dcdb0de |
| institution | Kabale University |
| issn | 2996-9409 |
| language | English |
| publishDate | 2024-08-01 |
| publisher | Open Exploration Publishing Inc. |
| record_format | Article |
| series | Exploration of Digital Health Technologies |
| spelling | doaj-art-41e8b358a71b4f8abbb07faf5dcdb0de2025-08-20T03:24:43ZengOpen Exploration Publishing Inc.Exploration of Digital Health Technologies2996-94092024-08-012420221710.37349/edht.2024.00022Developing a multi-variate prediction model for COVID-19 from crowd-sourced respiratory voice dataYuyang Yan0https://orcid.org/0009-0005-9914-1480Wafaa Aljbawi1Sami O. Simons2https://orcid.org/0000-0002-4296-5076Visara Urovi3https://orcid.org/0000-0003-2817-3950Institute of Data Science, Maastricht University, 6229 EN Maastricht, The NetherlandsInstitute of Data Science, Maastricht University, 6229 EN Maastricht, The NetherlandsDepartment of Respiratory Medicine, Maastricht University Medical Center, Maastricht University, 6229 HX Maastricht, The NetherlandsInstitute of Data Science, Maastricht University, 6229 EN Maastricht, The NetherlandsAim: COVID-19 has affected more than 223 countries worldwide and in the post-COVID era, there is a pressing need for non-invasive, low-cost, and highly scalable solutions to detect COVID-19. This study focuses on the analysis of voice features and machine learning models in the automatic detection of COVID-19. Methods: We develop a deep learning model to identify COVID-19 from voice recording data. The novelty of this work is in the development of deep learning models for COVID-19 identification from only voice recordings. We use the Cambridge COVID-19 Sound database which contains 893 speech samples, crowd-sourced from 4,352 participants via a COVID-19 Sounds app. Voice features including Mel-spectrograms and Mel-frequency cepstral coefficients (MFCC) and convolutional neural network (CNN) Encoder features are extracted. Based on the voice data, we develop deep learning classification models to detect COVID-19 cases. These models include long short-term memory (LSTM), CNN and Hidden-Unit BERT (HuBERT). Results: We compare their predictive power to baseline machine learning models. HuBERT achieves the highest accuracy of 86% and the highest AUC of 0.93. Conclusions: The results achieved with the proposed models suggest promising results in COVID-19 diagnosis from voice recordings when compared to the results obtained from the state-of-the-art.https://www.explorationpub.com/uploads/Article/A101122/101122.pdfcovid-19 diagnosisvoice analysismachine learningdeep learningmel-spectrogrammfcc |
| spellingShingle | Yuyang Yan Wafaa Aljbawi Sami O. Simons Visara Urovi Developing a multi-variate prediction model for COVID-19 from crowd-sourced respiratory voice data Exploration of Digital Health Technologies covid-19 diagnosis voice analysis machine learning deep learning mel-spectrogram mfcc |
| title | Developing a multi-variate prediction model for COVID-19 from crowd-sourced respiratory voice data |
| title_full | Developing a multi-variate prediction model for COVID-19 from crowd-sourced respiratory voice data |
| title_fullStr | Developing a multi-variate prediction model for COVID-19 from crowd-sourced respiratory voice data |
| title_full_unstemmed | Developing a multi-variate prediction model for COVID-19 from crowd-sourced respiratory voice data |
| title_short | Developing a multi-variate prediction model for COVID-19 from crowd-sourced respiratory voice data |
| title_sort | developing a multi variate prediction model for covid 19 from crowd sourced respiratory voice data |
| topic | covid-19 diagnosis voice analysis machine learning deep learning mel-spectrogram mfcc |
| url | https://www.explorationpub.com/uploads/Article/A101122/101122.pdf |
| work_keys_str_mv | AT yuyangyan developingamultivariatepredictionmodelforcovid19fromcrowdsourcedrespiratoryvoicedata AT wafaaaljbawi developingamultivariatepredictionmodelforcovid19fromcrowdsourcedrespiratoryvoicedata AT samiosimons developingamultivariatepredictionmodelforcovid19fromcrowdsourcedrespiratoryvoicedata AT visaraurovi developingamultivariatepredictionmodelforcovid19fromcrowdsourcedrespiratoryvoicedata |