Database Reference
In-Depth Information
(Java et al., 2007). Further, existing and new ap-
plications will incorporate personalization and
context-awareness features, thus meshing with
the daily reality of people. Thus, data will be
recorded that will capture all aspects of people's
lives, such that new insights into human behavior
will be possible through “reality mining” (Eagle
& Pentland, 2006), raising new concerns about
privacy. By closely monitoring the actions of the
individual, machine learning will be employed to
predict future consumption patterns, e.g. in order to
enable effective personalized advertising schemes
(Piwowarski & Zaragoza, 2007) or to spawn social
interactions between strangers based on profile
matching (Eagle & Pentland, 2005).
There is already evidence that the analysis
of data coming from social bookmarking usage
can be beneficial for search engine tasks, such as
new website discovery and authoritative online
resource identification (Heymann et al., 2008).
In the long run, advanced information extraction
and semantic analysis techniques, e.g. automatic
quality evaluation of user contributed content (Hu
et al., 2007), are expected to be deployed in real-
world applications (e.g. Wikipedia) and provide
the basis for even more advanced collaborative
applications, such as innovation management
platforms (Perlich et al., 2007). The require-
ments of such applications will in turn instigate
research into underlying technology disciplines,
i.e. database systems for the support of online
community services and collaborative platforms
(Ramakrishnan, 2007), as well as frameworks for
scalable knowledge discovery from streaming data
(Faloutsos et al., 2007).
resources submitted by other users as well as to
form relations with each other has stimulated
intensive user activity in SBS, such that massive
amounts of web activity data capturing the content
consumption patterns of users are produced in a
streaming fashion. This novel content consump-
tion paradigm has spawned a series of interest-
ing research questions related to the generation
and evolution of such patterns. These questions
pertain to the distributional attributes of online
resource popularity, the temporal patterns of
content consumption by users, the semantic as
well as the social factors affecting the behavior
of the masses with regard to their preferences for
online resources.
This chapter presented an overview of exist-
ing research efforts that are germane to these
questions and provided additional insights into
the phenomena taking place in the context of an
SBS by carrying out an analysis of a large dataset
collected from Digg. The power-law nature of web
resource popularity was established in accordance
with previous studies of similar online systems
(Cha et al; 2007, Hotho et al., 2006a; Halpin et al.,
2007). Furthermore, a set of characteristic tempo-
ral patterns of content consumption were revealed
which confirmed previous findings about social
media content popularity evolution (Lerman;
2007). In addition, a preliminary investigation
into the semantic elements of content popularity
lent support to the hypothesis that popularity is
affected by the semantic content and the linguistic
style. What is more, it was empirically shown that
users of SBS are socially susceptible, i.e. they
tend to express interest for online resources that
are also considered interesting by their online
“friends”.
Finally, the chapter provided an outlook on
the exciting new prospects for online content
publishing and mining on the massive amounts of
data produced in the context of SBS and related
applications.
concluSIon
The widespread adoption of SBS has transformed
online content consumption due to the powerful
features that such systems offer to their users. The
possibility for users to submit links to content of
their interest, tag, rate and comment on online
Search WWH ::
Custom Search