PNNL

The April 2018 issue of Scientific American features work led by Svitlana Volkova, a scientist with the National Security Directorate’s Computing and Analytics Division, on using social media to predict influenza outbreaks. In the article, “#Flu: Mining Social Media to Predict Outbreaks,” science writer Rachel Berkowitz compared the team’s work with previous research about flu outbreak prediction, which searched social media for illness-related words.

In their work, Volkova and her team used deep learning to correlate non-illness-related words in Twitter feeds with patterns of flu outbreaks. The algorithm “learned” to predict outbreaks up to two weeks in advance accurately. Such prediction may enable officials to take proactive public health measures that could save lives. After PLOS ONE published the article by Volkova, Ellyn Ayton, Katie Porterfield, and Court Corley last December, Berkowitz (who visited PNNL in 2012) wrote the Public Health section article for Scientific American.

Reference:

Volkova S, E Ayton, K Porterfield, and CD Corley. 2017. “Forecasting influenza-like illness dynamics for military populations using neural networks and social media” PLOS ONE 12(12):e0188941. DOI: 10.1371/journal.pone.0188941.