Database Reference
In-Depth Information
such interactions do have a significant influence over individuals, which
may propagate in the social network over time. Such an approach [54,
139, 140] has also been applied to the problem of geriatric care. This
is because medical conditions such as dementia in older patients show
up as specific kinds of activity patterns over time. It has been shown in
[54], how such activity recognition methods can be used in the context
of geriatric care.
From a predictive modeling perspective, a key challenge which arises
is that a large amount of data may potentially need to be collected simul-
taneously from a large number of patients in order to make accurate real
time predictions. This requires the design of fast data stream processing
algorithms [7]. A recent paper [89] proposes a number of real-time data
stream mining methods for fast and effective predictive modeling from
sensor data. This kind of approach can be used for a wide variety of
medical conditions, though the nature of the data collected and the pre-
dictive modeling would depend upon the nature of the disease modeling
at hand. For example, the work in [84] discusses a variety of methods
which can be used for diabetes monitoring with the use of collected data.
Another interesting method for health and fitness monitoring has been
developed in [119], in which modern mobile phones are used in order to
both sense and classify the activities of an individual in real time. It
has been shown that such machine learning algorithms can be used in
conjunction with the collected data in order to provide effective moni-
toring and feedback. A discussion of some of the challenges in selecting
sensors for health monitoring with the use of participatory sensing may
be found in [41].
9. Future Challenges and Research Directions
In this chapter, we examined the emerging area of integrating sensors
and social networks. Such applications have become more commonplace
in recent years because of new technologies which allow the embedding
of small and unobtrusive sensors in clothing. The main challenges of
using such technologies are as follows:
Such applications are often implemented on a very large scale. In
such cases, the database scalability issues continue to be a chal-
lenge. While new advances in stream processing have encouraged
the development of effective techniques for data compression and
mining, mobile applications continue to be a challenge because of
the fact that both the number of streams and rate of data collection
may be extremely large.
Search WWH ::




Custom Search