Biology Reference
In-Depth Information
15.5 Future Work
Now that the world has past the initial stages of the A (H1N1) epidemic, people
have begun asking, “Given the current state of preparedness, why didn't [we]
detect it earlier?” Mexican authorities detected the epidemic as early as March
18, but considered it to be late season influenza (Wikipedia 2009). Moreover, a
May 11 study in the journal Science estimated that there could have potentially
been as many as 23,000 suspected cases in Mexico before the 23rd of April,
with possibilities of concession possibly beginning as early as mid-February
2009. The linguistic signals before April 21, 2009, need careful examination, but
are most likely to be found in the Spanish-language media, which BioCaster,
unfortunately, did not cover at the time. The clear lesson (Figure 15.6) is the
advantage of having multilingual coverage. Work in BioCaster has already
begun to expand its language sources to several major world languages, includ-
ing Chinese, English, French, Russian, Spanish, and Portuguese.
Electronic publication has removed the physical space limits that previ-
ously existed in newspapers; modern newswire reports circulate widely and
are republished in many different forms. The disparity between real-world
frequencies of events and news frequencies needs to be handled to allow for
better situation awareness. How to normalize the frequency and volume of
news so as to reduce the effects of reporting bias also requires further study.
Understanding location and time are key foundations to effective GHIS.
Although studies have taken place in the area of temporal understanding
(e.g., TimeML 6 ), few guidelines or resources exist for geolocating events
outside of the English language. The GeoNames database 7 is one example
10000 km
Figure 15.6
Distribution of influenza reports found by BioCaster in the open media for April 21, 2009, to
June 1, 2009 (1≤ n≤1176). In comparison to WHO situation report case count totals for June 1, the
plot indicates under-reporting in South America, several false positives in Western Africa, and
under-reporting in Spain and South America and over-reporting in India and Western Africa.
Outbreaks of A (H5N1) occurring simultaneously were also sometimes not easily differenti-
ated from A (H1N1) as seen in the Egypt cluster. Maps are generated by GPSVisualizer
Search WWH ::




Custom Search