Geography Reference
In-Depth Information
attributes, the value observed at a location is expected to be similar to that of its
neighboring locations. The spatial locations where attribute value are significantly
different from surroundings are called spatial outliers. For example, a huge new
“McMansion” (
Wikipedia 2012-M
) built in an older modest neighborhood is
considered a spatial outlier. Finding outliers in spatial datasets is an important
activity in application domains that use image analysis or spatial statistic techniques.
17.4.2
Interval and Boundary Changes
A change pattern can also appear as a linear collection of locations/time frames.
Two specific patterns of this type are change interval and boundary.
17.4.2.1
Interval Changes
On the temporal dimension of a dataset, an increase of value may occur within
an interval of time where a persistent trend can be observed. A large fraction of
time frames in this period may experience a significant shift of value. Figure
17.9
a,
shows intervals of change in the smoothed Sahel rainfall index time series, where
the precipitation decreased rapidly between 1966 and 1971.
Fig. 17.9
An example of ST change interval pattern (
a
) An example of change interval in a time
series (
b
) An example of change interval in a spatial path
Search WWH ::
Custom Search