Geography Reference
In-Depth Information
on April 15th, 2012. The spatiotemporal pattern of terms is used to understand
more about the dynamic tornado- related events across space. The map reveals
that
large hail, lightning, heavy downpour,
and
strong thunderstorm
were associated
with parts of Texas, Louisiana, Wisconsin, Michigan and Minnesota. Facilities were
impacted by tornadoes, for example,
Wichita airport was damaged
in Kansas. In
addition, relief activities were begun in Nebraska, Kansas, and Oklahoma. For
example,
trucks were being sent out
for rescue in Oklahoma. Making sense of
spatiotemporal information through extracted semantic information reveals details
about the dynamic aspects of hazards for users.
The results can also be presented at a higher level of abstraction (Fig.
10.7
).
In this case, terms are presented at the granularity of super classes:
Wa r n i n g ,
MeteorologicalPhenomena, ClimaticHazard, HazardThreat, Response
and
Recov-
ery,
based on the hazard ontology. For example,
lightning, downpour, hail,
and
storm
are aggregated to the class:
MeteorologicalPhenomena
while
twister
and
tornado
are aggregated to the class of
ClimaticHazard.
It is possible to follow the sequence
of these higher level semantics of a hazard (e.g., hurricane, response, recovery)
automatically based on analysis of text. The map shows how the five classes vary
spatially as per the report on April 15th, 2012. Based on this representation and for
this document date, it is possible to see which areas are still experiencing storms
(i.e., meteorological phenomena) and where response is beginning to occur.
Fig. 10.7
Semantics extracted from a news report about tornadoes that struck Oklahoma, US on
April 15, 2012
Search WWH ::
Custom Search