Information Technology Reference
In-Depth Information
If
pd
y
(
p, p'
)
Į
and
pd
y
(
p',p
)
Į
, then we define a bidirectional edge
between person
p
and
p'
. If
pd
y
(
p, p'
)
į
and
pd
y
(
p',p
)
ȝ
, then we define
an edge from person
p
to person
p'
.
2
The size and colour of a node represent the importance of a person.
The size of node is defined by an attraction power of person
p
described
by
p''
א
p
pd
y
(
p'',p
), which means the summation of other people's person
dependency on person
p
. The colour of the node is defined by the ratio of
the number of in-links to out-links of the node. We use the following
colours as default colours that can be modified by the visualization system
described in Section 6.1.
y
#In Link > #Out Link
: dark purple
y
#Out Link > #In Link
: light purple
y
#In Link = #Out Link
: gray
The length of an edge represents the strength of a relationship between
two people. It is defined by the difference in person dependencies
pd
y
(
p,p'
)
-
pd
y
(
p',p
). If person dependency is strong, the edge shortens. If one person
one-sidedly depends on another, the edge lengthens.
5 Extracting Characteristics of Relationships between
Historical Figures
Once the network of historical figures (persons) is generated, the next
stage is to extract keywords in events related to the two historical figures
to define the characteristics of their relationship. We then add such
keywords as labels to their relationships and attach colour to every label.
This helps users to investigate characteristics of relationships between two
historical figures.
The number of keywords for the case study is 3034, for a period of 20
years (from 1560 to 1580). With such a large number of keywords, it is not
practical to use colour as the distinguishing factor. Moreover, it is difficult
to understand relationships between keywords. It is worth noting that most
keywords are related to conflicts or battles, and the rest are related to
peace and collaboration.
To address this problem, we first extract clusters of keywords on the
basis of co-occurring keywords. We then extract common clusters of
keywords related to two people for every year. Such common clusters of
keywords represent components of the relationship between people.
2. We use
Į
=1 and
į
=
ȝ
=0.8 in this chapter.
Search WWH ::
Custom Search