what-when-how
In Depth Tutorials and Information
statistical surveys. However, with the advent of combination tests, this is much less
of a predicament at the moment. An ongoing problem is the shortage of enough
computing resources to deal with large datasets. It is frequently a challenge to con-
strict a social network. If we are looking at needle-sharing among drug users, we
can artificially compress the network at some arbitrary border line, such as an urban
center or district, but this would misrepresent the information. Yet we should not
let the network get so large that it becomes unmanageable.
16.4 MiningHiddenCommunitiesinHeterogeneous
SocialNetworks[1]
It is essential to understand and explore the concept of community mining in
more complex and diverse social networks. On account of its rapid growth, the
social network becomes a center of interest for most people, especially researchers
who want to know what leads people to join a social network and what makes
a social network active and worldwide in reach. In fact, community mining is
an important tool that will help us to understand those parameters. Obviously,
assuming that the social network is unirelational and “mining results are indepen-
dent of the user's needs or preferences” will leave out precious hidden community
information.
Nowadays, there are a huge number of social networks, and it is certain that
each network provides users some features that others do not have; moreover, a
client
A
may have a distinct relationship with clients
B
and client
C
, and user
C
may also have a different relationship with user
D
and the same relationship
with
A
or
F
. his means that each user creates a small network inside the colossal
social network, depending on his or her interests, desires, and favorites. herefore,
the new approach to analyzing social network should include all those factors
since, as stated earlier, the social network itself is becoming a multirelational social
network.
As stated earlier, one of the main focuses here is to determine some criteria
(maybe formulas) that will allow us to know whether a social network is strong
or not. Because there are different relations between users in the social network, it
might be wise to check the strength and importance of those relations. his leads
researchers to recommend a mathematical model for relation extraction and selec-
tion. he main idea of this algorithm is to model the problem as an optimization
problem, which means that once the algorithm is built, it can be applied to any
social network to know how influential it is.
For this approach, each relation will be represented by a graph with a weight
matrix, and each element of the matrix illustrates its relation strength between
the two corresponding objects. In order to make the algorithm more efficient,
researchers added the feature extraction problem to the relation extraction prob-
lem. In fact, the feature extraction problem is a way of reducing the dimensions of
