A new algorithm mines data from Wikipedia to track flu cases across the United States.
The program is designed to monitor certain entries that a sick person would look up, such as "flu season" and "fever," and hourly download publicly available information on how many people nationwide accessed the pages. In comparing their data with figures from the U.S. Centers for Disease Control, the researchers found they could accurately predict the number of cases in the county two weeks earlier and with a difference of just 0.27 percent.
In addition, tweets about sickness, mentions of activities one might need to be healthy, and changes in Twitter use could be useful for monitoring a specific group of people, says Pennsylvania State University's Todd Bodnar. His team at the Center for Infectious Disease Dynamics analyzed the Twitter feeds of 104 students, and its new algorithm was able to identify with 99 percent accuracy if a student had suffered from flu during a given month.
From New Scientist
View Full Article
Abstracts Copyright © 2014 Information Inc., Bethesda, Maryland, USA