Timely text mining comes to the aid of 'swine flu' fighters
Jun 16, 2009 at 08:19 pm by
Staff
Systems software developer Nstein says its text mining engine is playing a part in the early warning of swine flu, with the Global Public Health Intelligence Network relying on its semantic technology.
The company says GPHIN Centre for Emergency Preparedness and Response Public Health Agency of Canada chief Abla Mawudeku was in a meeting at the World Health Organization (WHO) in Geneva, Switzerland, discussing how new technologies were assisting in determining the risks and threats to humans around the globe – when the H1N1 virus first appeared on the world stage.
The ‘swine flu’ virus was being tracked by analysts who were ever-vigilant of the progression of a new risk. It was while at this meeting that Mawudeku's team sounded the alarm.
Mawudeku’s team uses the latest technology to spot threats to human life. Every day multilingual analysts comb through content from multiple sources including news aggregators such as Factiva, which has been parsed for about 1000 concepts such as “mysterious ailments” and “outbreak.”
The GPHIN system uses Nstein's TME to assign relevancy scores to numerous articles coming from worldwide sources, and assigns a relevancy score to each article. TME also de-dupes redundant news articles. The system helps analysts count and track instances of possible threats. It is precisely this monitoring by organizations like GPHIN that triggers responses – such as WHO declaring H1N1 a geographic pandemic, which in turn hurries the development of vaccines.
A recent National Post article cited GPHIN's importance to detecting global threats (http://www.nationalpost.com/news/canada/story.html?id=1562873). Unlike its American counterpart ProMed, GPHIN does not just track diseases. The group monitors any threat to human life – natural or man-made: bush fires in California, pestilence outbreaks in Africa, even theft of nuclear elements. As such the Canadian team and its technology are monitoring more than 1000 types of human harms, around the globe. The team publishes eight different reports, three times a day.
It is Mawudeku's group and the GPHIN system along with the American ProMed that assisted World Health Organisation scientists in determining if an outbreak was contained or reaching pandemic.
“The job is a tremendously stressful one,” said Mawudeku. “We are normally adding 4,000 articles a day. Right now (during the influenza outbreak), we are overwhelmed with more than 20,000 a day. It would be impossible to track this volume without technology – which is exactly what we were discussing when the alert came out.”
