Database Reference
In-Depth Information
it is possible to determine the key events from the underlying social
network from the patterns in the underlying text stream [11]. Other
general methods for event detection in streams are discussed in [3, 65,
69, 76].
4.2 Cosmological Applications
In recent years, cosmological applications have created large volumes
of data. The installation of large space stations, space telescopes and ob-
servatories result in large streams of data on different stars and clusters
of galaxies. This data can be used in order to mine useful information
about the behavior of different cosmological objects. Similarly, rovers
and sensors on a planet or asteroid may send large amounts of image,
video or audio data. In many cases, it may not be possible to manually
monitor such data continuously. In such cases, it may be desirable to use
stream mining techniques in order to detect the important underlying
properties.
The amount of data received in a single day in such applications can
often exceed several tera-bytes. These data sources are especially chal-
lenging since the underlying applications may be spatial in nature. In
such cases, an attempt to compress the data using standard synopsis
techniques may lose the structure of the underlying data. Furthermore,
the data may often contain imprecision in measurements. Such impre-
cisions may result in the need for techniques which leverage the uncer-
tainty information in the data in order to improve the accuracy of the
underlying results.
4.3 Mobile Applications
Recently, new technologies have emerged which have allowed the con-
struction of
wearable sensors
in the context of a variety of applications.
For example, mobile phones carry a wide variety of sensors which can
continuously transmit data that can be used for
social sensing
applica-
tions [62]. Similarly, wearable sensors have been designed for continuous
monitoring in a wide variety of domains such as health-care [46, 71] or ve-
hicular participatory sensing [47]. All vehicles which have been designed
since the mid-nineties carry an
OBD Diagnostic System
, which collects
a huge amount of information from the underlying vehicle operation. It
is possible to use the information gleaned from on-board sensors in a
vehicle in order to monitor the diagnostic health of the vehicle as well as
driver characterization. Another well known method is the
VEDAS
sys-
tem [55], and the most well known commercialized system is the
OnStar
system designed by General Motors. Such systems require quick analysis
