Harmonization Challenges in Data Collection of COVID-19 Misinformation
| Main Authors: | Grant, Joshua N., Tansakul, Varisara, Eaton, Bryan, Thakur, Gautam, Smyth, Martin, Smith, Monica |
|---|---|
| Format: | Proceeding Journal |
| Bahasa: | eng |
| Terbitan: |
, 2021
|
| Subjects: | |
| Online Access: |
https://zenodo.org/record/4705303 |
Daftar Isi:
- Misinformation surrounding COVID-19 poses a serious threat to human life. The collection, categorization, and harmonization of misinformation are crucially needed to understand the impacts of the distribution of false or misleading information on public safety and human health. Identifying data sources to collect misinformation requires initial review to evaluate if the fact-checking website reports fact-checking information in useful manners for downstream harmonization and analysis. Fact-checking websites have dissident classification methods, requiring a mapping categorization step in the data flow process. The harmonization of collected information allows for the creation of a rich dataset allowing for insights to be drawn from multiple fact-checking websites, such as understanding the spread of misinformation around the world. The collected information from websites has various data schemas regarding claims and original sources of the (mis)information. Due to this challenge, manual effort is required for reviewing and determining the appropriate categories for the unharmonized data. The categorization and harmonization challenges become apparent as more data from more websites are collected.