Application of text data mining in covid-۱۹: a systematic review

Background and aims: The Covid-۱۹ outbreak has led to an unprecedented amount of textualdata being generated from various sources such as scientific databases, websites, and social mediaplatforms worldwide. Hence, automatic methods of analyzing textual data have become increasinglyimportant. Therefore, the aim of the study is investigate the applications of text data miningduring the Covid-۱۹ pandemic.Methods: This study follows the PRISMA guideline for a systematic review. The data for thisreview was obtained through a search of PubMed, using relevant keywords such as “COVID-۱۹,”“Text Mining,” and “Text Data Mining” and their associated synonyms, from the beginning of thepandemic until December ۴th, ۲۰۲۲. Original articles published in English that used of text datamining in the methodology of the research were considered for this review.Results: After conducting a thorough search and applying relevant inclusion criteria, a total of۱۱۷ articles were selected out of the initial ۱۹۰ identified. The majority of the studies includedwere published in ۲۰۲۲, indicating an increasing interest in the application of text mining duringthe pandemic. Among the various sources of textual data analyzed, social media posts and commentswere the most commonly studied (۴۹ studies), followed by scientific literature and articles(۲۲ studies), and questionnaires and interviews (۱۸ studies). The primary tools used for text analysiswere the Python programming language (۵۲ studies), R software (۱۹ studies), and KH Codersoftware (۶ studies). In terms of algorithms, topic modeling techniques such as LDA, STM, andLSA were widely used, as well as sentiment analysis methods such as BERT and VADER.Conclusion: The results have shown that text mining techniques have been widely used in theanalysis of textual data in the Covid-۱۹ pandemic. In addition, the results obtained through textmining can assist health researchers and policymakers in identifying thematic trends and publicationpatterns across a range of texts, including scientific articles, debates, and public opinions.This can be instrumental in informing evidence-based decision-making during times of crisis.

کلیدواژه ها:

COVID-۱۹ ، Text Mining ، Text Data Mining ، Data Mining