Discovering Time-Varying Public Interest for COVID-19 Case Prediction in South Korea Using Search Engine Queries: Infodemiology Study
2024

Predicting COVID-19 Cases Using Search Engine Queries

publication 10 minutes Evidence: high

Author Information

Author(s): Mavragani Amaryllis, Chun June Young, Yang Shihao, Ahn Seong-Ho, Yim Kwangil, Won Hyun-Sik, Kim Kang-Min, Jeong Dong-Hwa

Primary Institution: The Catholic University of Korea

Hypothesis

Can search engine queries effectively predict COVID-19 case numbers by reflecting public interest over time?

Conclusion

The study developed a model that predicts COVID-19 cases by analyzing search engine queries, outperforming previous methods.

Supporting Evidence

  • The model outperformed previous methods that relied on static symptom-based queries.
  • Search engine queries showed a high correlation with confirmed COVID-19 cases.
  • Temporal variations in public interest were effectively captured by the model.

Takeaway

This study shows that looking at what people search online can help predict how many COVID-19 cases there will be in the future.

Methodology

The study used word embedding models to analyze search queries and applied ElasticNet regression for predictions.

Potential Biases

The reliance on news articles may introduce bias based on the selection of media outlets.

Limitations

The model's performance may decline as public interest in COVID-19 decreases over time.

Statistical Information

P-Value

P=.02

Statistical Significance

p<0.01

Digital Object Identifier (DOI)

10.2196/63476

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication