Search Here

Every year, health officials tally up the total number of flu cases to see how severe the season was. However, oftentimes, influenza cases go underreported. In order to more accurately predict the effects of the flu season, researchers believe that using Wikipedia to track people's searches on the flu can be helpful.

For this study, the research team developed a new data-analysis system. The system tracks how often people visited Wikipedia websites that were related to the flu between December 2007 and August 2013, which encompassed six flu seasons. To test the effects of the system, the researchers compiled estimates two weeks before official data were released by the U.S. Centers for Disease Control and Prevention (CDC). The system was capable of accurately predicting the peak flu week numbers for three out of the six seasons.

"We were able to get really nice estimates of what the [flu] level is in the population," said study author David McIver, a postdoctoral fellow at Boston Children's Hospital reported by FOX News.

The system works by focusing on key terms related to flulike illnesses. The researchers also examined how often a particular article was viewed per hour. The Wikipedia system worked better than the Google-based system that only estimated two out of the six seasons correctly. The researchers stated that the Google Flu Trends system was prone to "overfitting," which meant that Google tended to include irrelevant searches about the flu to overall flu estimates. For example, during a flu pandemic, more people might Google search for information about the pandemic. However, that does not mean that they are experiencing flu symptoms. The Wikipedia-based system tries to weed out those types of searches.

The researchers stated that they did not create the system to try to replace the CDC. However, they hope that their system could help the CDC provide even more accurate estimates. The study was published in the journal, PLOS Computational Biology.