Utilizing Google Trends data for effective modeling of COVID-19 outcomes: a vector auto regression (VAR) approach

Over the past decade, especially during the three-year COVID-19 epidemic, infodemiology, which uses web-based data for public health issues, has helped assess and anticipate human behavior. Google's real-time population data can indicate popular interest during a pandemic. This study analyzes Google Trends (GT) data of COVID-19-related search phrases and CDC information to forecast US daily new cases, cumulative cases, and deaths. Seasonality and trends are removed from trend data using an Augmented Dickey-Fuller (ADF) test. Using an eight-week forecast horizon, Granger Causality tests and Vector Auto Regression (VAR) models predict COVID-19 cases and deaths. GT's relative search volume of COVID-19 keywords and CDC's daily confirmed cases and cumulative deaths determine VAR model input search terms. RMSE (root mean square error), MAE (Mean Absolute Error), MAPE (Mean Absolute Percentage Error), and MASE (Mean Absolute Scaled Error) were used to compare forecast accuracies. Also analyzed are Long-Covid search trends.

Grantor dc:publisher

University of Tennessee at Chattanooga

Author dc:creator

• Osei, Gertrude

Contributors dc:contributor

• Gao, Lani
• Ma, Ziwei; Barioli, Francesco
• College of Arts and Sciences

• Internet in medicine--United States
• COVID-19 Pandemic, 2020---Social aspects--United States
• Data mining

Licence

• http://rightsstatements.org/vocab/InC/1.0/

Language dc:language

English, eng

Repository record dc:identifier

https://scholar.utc.edu/theses/809

OAI identifier oai:identifier

oai:scholar.utc.edu:theses-1973