Automatic extraction of knowledge for diagnosing COVID-19 disease based on text mining techniques: A systematic review
In late December 2019, an epidemic of the novel coronavirus (COVID-19) was informed, and because of the quick diffusion of the infection in various regins of the world, the World Health Organization proclaimed an emergency. In this context, researchers are urged and encouraged to research in various...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
International University of Sarajevo
2021
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/97478/1/SitiSophiayatiYuhaniz2021_AutomaticExtractionOfKnowledgeForDiagnosingCOVID19.pdf http://eprints.utm.my/id/eprint/97478/ http://dx.doi.org/10.21533/pen.v9i2.1945 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | In late December 2019, an epidemic of the novel coronavirus (COVID-19) was informed, and because of the quick diffusion of the infection in various regins of the world, the World Health Organization proclaimed an emergency. In this context, researchers are urged and encouraged to research in various fields, to stop the spread of this deadly virus. To this end, we propose a systematic review that addresses the techniques and methods of artificial intelligence in diagnosing COVID-19 disease. The main aim of the current systematic review was to highlight the gaps and challenges within the academic literature of the disease COVID-19, which included the characteristics of the data, machine learning algorithms applied to the diagnosis of COVID-19, and using natural language processing (NLP)to reveal clinical data for COVID-19 disease.Seven reliable databases were used, namely Web of Science, ScienceDirect, IEEE Xplore, Scopus, PubMed, springer and google scholar, to obtain studies related to the specific topic many filtering and surveying stages were conducted consistent with the inclusion and exclusion criteria, to screen the acquired 1115 papers.We identified the bottleneck in explaining data as one of the major barriers to machine learning and NLP approaches. Supervised machine learning has been explored as an active method for diagnosing COVID-19 disease. Future studies in this area will benefit from alternatives like increasing the volume of data, using intelligence swarms to obtain accurate features, and using unsupervised learning that does not require explanatory data. Thus, this research supported us to get a more practical comprehension of the gaps and provide possible solutions for filling these gaps. |
|---|