Information Technology Reference
In-Depth Information
identify the habits of the residents in terms of presence. He proposes to employ
energy disaggregation techniques to make inference about the use of certain
household appliances, which indicate the physical presence of the occupants. This
approach is also referred to as non-intrusive load monitoring, which has the bene
t
that it refrains from using an additional sensor infrastructure.
Focusing on personal media consumption, Acar et al. present in Chap. 11 an
approach for identifying a certain type of pattern present in Hollywood movies or
user generated videos. This pattern is
These movies or videos contain
two modalities (audio and visual), each modality being directed to a different sense
of the media consumer (hearing and seeing), therefore allowing an
violence.
which does not actually exist, or is limited to textual information. Detecting violent
content in movies and videos is one application which neatly illustrates the meaning
of
immersion,
Acar et al. achieve this by extracting meaningful
features from the data, and by classifying those features using advanced machine
learning techniques. They also present a user interface designed to allow the con-
sumer to browse data and search for
bridging the semantic gap.
scenes.
The scenario that is outlined in Chap. 12 appears in the context of the automotive
industry, where the objective is to optimize vehicle engines with regard to exhaust
emission. Spiegel presents a data mining approach that was developed in cooper-
ation with researchers and engineers from one of the leading car manufacturers,
who aim to run emission simulations based on operational proles that characterize
recurring driving behavior. In order to obtain real-life operational pro
violent
les, the
automotive engineers collect sensor data from test-drives for various combinations
of driver, vehicle, and route. Such measurements can also be considered as high-
dimensional time series, where each dimension represents the progression of a
certain physical quantity, such as the engine temperature, during car drive. Spie-
gel
s proposed approach is able to identify time series representatives that best
comprehend the recurring temporal patterns contained in a corresponding dataset.
He applies this approach to determine operation pro
'
les that comprise frequently
recurring driving behavior patterns, but his introduced model can also be used for
time series datasets from other domains.
The
nal scenario of this section, presented in Chap. 13 , focuses on the related
topic of traf
c optimization. The ever-increasing urbanization has signi
cantly
aggravated the traf
c situation in megacities. Common travel habits, such as using a
vehicle, are challenged by factors like severe congestion,
insuf
cient parking
availabilities, or present
c entails an
increased level of noise and greenhouse gases and thus affects residents even more.
Acar et al. present an approach to utilize means of transportation in a more effective
and sustainable fashion in order to increase the quality of life in cities and to
contribute to global environmental objectives. They describe a travel assistance
system that proposes intermodal traveling options which are tailored to drivers
fuel prices. Furthermore, growing traf
'
needs. Different information channels are integrated in the system. One of these
channels is information derived from video analysis.
Search WWH ::




Custom Search