You are on page 1of 4

Review Paper

2012 ACEEE
DOI: 01.IJRTET.7.1.
Int. J. on Recent Trends in Engineering and Technology, Vol. 7, No. 1, July 2012
40
On Application of Rough Set and Neighborhood
Theory in Social Network
Anirban Mitra
1
, and Prasanta Padhi
2
1
Dept. Of CSE, MITS, Rayagada 765017, Odisha, India
2
Sysnet Global Technologies Private Limited, DRM-office, Samblpur, Odisha, India
Email: {mitra.anirban, prasanta.pdhi}@gmail.com
Abstract: Other than the user interface of any social network
site, there are lot many features, which are complex in nature,
makes the interface to work smoothly and efficiency. These
characteristic mainly works from behind of the interface. They
decrease the accessing time of users, increase reliability and
enhance resource access and management. These activities
can be broadly classified into three level of analysis:
clustering, mathematical modelling and data caching. The
problem of classification and clustering can be approach with
tools like Rough set. But, unless the social network is defined
in terms of equivalence relation, rough set cannot be
implemented directly. Hence we had defined the social network
by using the Fiksels concept. The concept of Neighborhood
theory may be termed as the extensions of rough set, which
includes topological spaces. Such tools also proves to be
efficient even, if the dataset consist of missing or uncertain
data. In our paper, apart from definitions and literature review
on Social network and Rough Set, we had established a
relationship between them. We have extended our discussion
for neighbourhood theory. The concept of neighbourhood
theory in social network can also be approached by using the
adjacency property for any non-directional graph under the
graph theory concepts. Further, we have discussed on the
relationship of rough set and neighborhood theory. Lastly, we
have observed that neighbourhood theory is having vast
applications in social network for problems like analysis of
node and clustering.
Index Terms: Social network, Societal Network, Rough Set
and Neighborhood Theory.
I. INTRODUCTION
The set of relationships between individuals, where each
individual is a social entity is known as Social network. Social
networks represent both a collection of ties between people
and strength of those ties. Mathematically, we can interpret a
social network as a social structure made of nodes (which are
generally individuals or organizations) that are tied by one or
more specific types of interdependency, such as values,
visions, ideas, financial exchange, friendship, sexual
relationships, kinship, dislike, conflict or trade [1,2,10].
Apart from the visual interface for a site that handles
such type of Network, there are lot many features, which are
complex in nature, makes the interface to work smoothly and
efficiency. Among the characteristic, that lies behind the
interface of such network is to decrease the access time of
users to assets, increase reliability and enhance resource
access and management. The working of these activities at
preliminary level may be broadly classifies as into three levels
of analysis: clustering, mathematical modelling and data
caching [3,4].
A rough set is represented by a pair of crisp sets, called
the lower and upper approximations of the set. The lower
approximation of a rough set comprises of those elements of
the universe, which can be said to belong to it definitely with
the available knowledge. The upper approximation on the
other hand comprises of those elements which are possibly
in the set with respect to the available information [6,7].
Joseph Fiksel [5] defined the term societal network instead
of social network to denote a mathematical system having
well defined properties, where node represent individual
people and arcs represent directed relations. Hence, societal
network is defined as labelled directed network
( , , ) N A T where ' ' N is a finite set of nodes,
' ' T
is a finite set of
relational type, and . A N T _ X The set
' ' A
represents all
the arcs in the network which link pairs of nodes. Thus, if
( , ; ) , x y u A e then there is an arc from node ' ' x to node ' ' y of
relational type ' '. u
II. ON ROUGH SET
The concept of rough set is an approach to deal with
imperfect knowledge. It was introduced by Z. Pawlak in 1982
[6,7,8]. Rough Set Theory (RST) can be approached as an
extension of the Classical Set Theory, for use when
representing incomplete knowledge. Rough sets can be
considered sets with fuzzy boundaries sets that cannot be
precisely characterized using the available set of attributes.
The basic concept of the Rough Set is the notion of
approximation space, which is an ordered pair
( , ), A U R = where : U nonempty set of objects, called
Universe and : R equivalence relation on , U called
indiscerinibility relations. If , x y U e and x R y then ' ' x and
' ' y are indistinguishable in
' ' A [6,7]
.
By a knowledge base, we express the a relational
system ( , ) A U R = , for any subset ( ) P R = c , t he
intersection of all equivalence relations in
' ' P
is denoted by
( ) IND P and is called the indiscernilibity relation over
' ' P
.
38
Review Paper
Int. J. on Recent Trends in Engineering and Technology, Vol. 7, No. 1, July 2012
2012 ACEEE
DOI: 01.IJRTET.7.1.
Given any X U _ and ( ) R IND A e , we derive two
s u b s e t s { / : } RX Y U R Y X = e _ a n d
{ / : } RX Y U R Y X = e = are called the R-lower and R-
upper approximations of X respectively. The R-boundary of
X is denoted by ( )
R
BN X and is given by
( )
R
BN X RX RX = . We say that
' ' X
is rough with respect
to
' ' R
if and only if RX RX = , equivalently ( )
R
BN X =
[9].
III. FIKSELS SOCIETAL NETWORK AND ROUGH SET
A societal network is defined to be a finite directed graph
in which individuals are represented by nodes, and relations
between individuals by labeled arcs. Each individual
undergoes state transition at discrete instant of time, so that
the societal network may be thought of as a deterministic
dynamic process [5]. It was shown by Fiksel that that the
individuals of such a network may be divided into
equivalence classes, so that the original network may be
represented by a reduced network containing one node for
each equivalence class. In other words, the concept of
structural equivalence and class structure can be used to
reduce a societal network to a network of equivalence classes
[5].
Definition: In a societal network, a pair of nodes x and y are
said to be structurally equivalent when,
(i) ( , ) ( , ) x y x y v = .
(ii) For any ( , ), ( , ) ( , ) x y R x R y e. =
(iii) If ( ), ( , ) ( , ) y G x R x y R y x e =
Where ( ) G x is the set of nodes adjacent to x .
Definition: A class structure of a societal network
( , , ) G N A T = is a partition of the set of nodes ' ' N into disjoint
sets
1 2 3
' , , ,... '
k
C C C C , called classes, when satisfies the
following two conditions : for any two nodes , x y of a class
' '
i
C there exist a mapping : ( ) ( ) M G x G y such that for
( ) : z G x e
(i) ( , ) ( , ( )) R x z R y M z =
and
(ii)
' ' z
is in the same class as ( ) M z .
Here,
' ' M
is the mapping and it may be one-to-many or many-
to-one.
The above definition of societal network clearly mention
on its properties of equivalence classes [5]. Hence, it is now
easy to relate the societal network in terms if reflexive,
transitive and symmetric. Equivalence relation is the
fundamental property for rough set based classification.
Therefore, rough set can now be used for clustering process,
once the societal network is defined in terms of equivalence
classes. Further, it is observed that the mentioned definition
of equivalence relation and equivalence classes from classical
set theory matches with the expression for lower
approximation and upper approximation of the rough set.
Definition for equivalence relation: A relation
' ' R
defined
on a set
' ' A
is said to be an equivalence relation in
' ' A
if and
only if
' ' R
is reflexive, symmetric and transitive.
Definition for equivalence classes: For any
' ' A
be a non
empty set,
' ' R
be an equivalence relation in ' '. A For each
, x A e the sets [ ] x are called equivalence classes of
' ' A
given by the relation
' ' R
defined as [ ] { | }. x y A y R x = e
IV. THEORY OF NEIGHBORHOOD SYSTEM
Neighborhood system was proposed by Lin [8]. The
concept of neighbourhood system discuss about nearness
or negligible distance. The definition of neighborhood system
includes rough set and topological space, and hence is termed
as extension of rough set.
A neighborhood system assigns each object a family of
non-empty subsets. Such subsets, called neighborhoods,
represent the semantics of near. The definition of
neighbourhood system according to Lin [8] is:
Let U be the universe of discourse and p be an object
in U.
1. A neighborhood, denoted by N(p), or simply N, of p
is a non-empty subset of U, which may or may not contain
the object p. A neighborhood system of an object p,
denoted by NS(p), is a maximal family of neighborhoods of
p. If p has no (non-empty) neighborhood, then NS(p) is an
empty family; in this case, we simply say that p has no
neighborhood.
2. A neighborhood system of U, denoted by NS(U) is the
collection of NS(p) for all p in U. For simplicity a set U
together with NS(U) is called a neighborhood system space
(NS-space) or simply neighborhood system.
3. A subset X of U is open if for every object p in X, there
is a neighborhood ( ) . N p X _ A subset X is closed if its
complement is open.
4. NS(p) and NS(U) are open if every neighborhood is
open. NS(U) is topological, if NS(U) are open and U is the
usual topological space. In such a case both NS(U) and the
collection of open sets are called topology.
5. An object p is a limit point of a set E, if every
neighborhood of p contains a point of E other than p. The set
of all limit points of E is call derived set. E together with its
derived set is a closed set.
6. NS(U) is discrete, if NS(U)= P(U), the power set.
7. NS(U) is indiscrete, if NS(U)=(U).
8. NS(U) is serial, if , ( ) p N p is non-empty..
9. NS(U) is reflexive, if , ( ) p p N p e .
10. NS(U) is symmetric, if
, , ( ) ( ) p q q N p p N q e .
11. NS(U) is transitive, if , , , , ( ) p q p r q N p e and
40
39
Review Paper
2012 ACEEE
DOI: 01.IJRTET.7.1.
Int. J. on Recent Trends in Engineering and Technology, Vol. 7, No. 1, July 2012
( ) ( ) r N q r N p e e .
12. NS(U) is Euclidean, if ( ) q N p e and
( ) ( ) r N p r N q e e .
V. RELATIONSHIP BETWEEN ROUGH SET AND NEIGHBORHOOD
SYSTEM
Lin had proposed the concept of neighbourhood system
[8]. It is the extension of the Rough Set. The pair ( ( ), ( )) R X R X
as pair of sets (no knowledge of R) does not characterize
Rough set. In general, for a subset X there is possible a
subset Y such that ( ) ( ), R X Y R X _ _

but ( ) ( ) R X R Y =

or
/ and ( ) ( ). R X R Y =
Let X be a subset of U. I[X] = {p : there is a ( ) N p X _ }
= interior of X, i.e., I[X] is the largest open set contained in X.
[ ] {} : ( ), C X p N p =
( ) X N p = } = closure of X i.e.,
C[X] is the smallest closed set contains X.
I[X] and C[X] are the lower and upper approximation in rough
set theory.
For R be an equivalence relation on U, it partitions U into
equivalence classes which can be denoted as R. This partition
can be regarded as neighbourhood system on U.
VI. DEFINING NEIGHBORHOOD THEORY IN SOCIAL NETWORK
USING GRAPH THEORETIC CONCEPTS
The concept of neighborhood was proposed by Everett,
Boyd and Borgatti [4] in 1990. According to them, for any
non-directional graph, which leads to non directional relations,
regular equivalence can be achieved. The neighborhood of a
node i in a graph consist of all nodes adjacent to node i. If
nodes i and j are regular equivalent, then for any node k
adjacent to node i there must be some node l adjacent to
node j and k and l must be regular equivalent. Since the
neighborhood of a node consist of all the nodes adjacent to
that node, nodes that are regularly equivalent must have
same equivalence classes of the node in their neighborhood
across all relations. Thus, in order to be regularly equivalent,
actors must be adjacent to the same kinds (equivalence
classes) of the other actors. This approach to defining regular
equivalence is especially useful for studying regular
equivalence in non directional relations. However, this
relationship does not hold fair for directional graph. We can
explain the scenario using the following two cases. The first
situation where a mother loves her baby and the baby equally
loves her mother can be explained in terms of non-directional
node. But it fails for situation like mother hold responsibilities
for her baby but baby is still young enough to understand
the responsibilities. In the second case, the relationship of
responsibility must be shown with the directed edge and the
graph becomes a directed graph.
40
VII. OBSERVATION
Once, the societal network is defined through its
properties of equivalence classes. It becomes easy to relate
the societal network with rough set. Lins definition of
neighborhood system is based upon Rough set. And rough
set is based on the equivalence relations. To analyze the
partition in societal network, on the basic of rough set, we
had defined the network through Fiksels definition. Again
Everett, Boyd and Borgatti had proposed the concept of
neighborhood on the basic of adjacency property of graph
theory. Their analysis of equivalence classes in graph also
leads to a kind of equivalence relation. Both the approach of
defining the relationship in social network is different. And
both the approaches have some restrictions. But, the analysis
shows that both concepts allow us to explore the problem of
clustering and classification in social network domain using
rough set.
CONCLUSION
In this paper, we have discussed on one of the challenge
of social network. We have broadly classified the activities
of social network at preliminary level into three level of
analysis: clustering, mathematical modelling and data caching.
Our research work deals with the problem of classification
and clustering which can be approach with tools like Rough
set. Further studying various extensions made in the field of
rough set, we have analyzed the concept of neighbourhood
system. Fiksel societal network also holds true for
equivalence relation. Hence using the Fiksels notation, rough
set and neighbourhood concept can be implemented in social
network problem. Further, Everett, Boyd and Borgatti
proposed the concept of neighbourhood using the concept
of node adjacency. Their approach also leads to equivalence
classes. Thus, with these two different approaches, it is true
that the clustering and classification problem in social
network can be handled with neighbourhood concept. Further,
as an extension of rough set, analysis of missing and
uncertain data set within the social network is also possible.
REFERENCES
[1] Backstrom L., Huttenlocher D., Kleinberg J. & Lan X., Group
formation in large social networks: Membership, growth, and
evolution, Proceedings of 12th Int. Conf. on KDD in Data
Mining, New York: ACM Press, 2006, Pp. 44 54.
[2] Danah Boyd and Jeff Potter (2003). Social Network Fragments:
An Interactive Tool for Exploring Digital Social Connections
Sketch at SIGGRAPH 2003. San Diego, California: ACM,
July 27-31, 2003.
[3] Danah Boyd, & Ellison N. B. Social network sites: Definition,
history, and scholarship Journal of Computer-Mediated
Communication, Vol. 13(1), 2007, Article 11-15.
[4] Everett M. G., Boyd J. P., Borgatti S. P., Ego-centered and
local roles: A graph theoretic approach, Jour. of Math.
Sociology, 15(3-4), pp. 163-172.
[5] Joseph Fiksel, Dynamic Evolution in Social Networks,
Jour. of Math. Sociology, 7(1980), 27-46.
40
Review Paper
Int. J. on Recent Trends in Engineering and Technology, Vol. 7, No. 1, July 2012
2012 ACEEE
DOI: 01.IJRTET.7.1.
[6] Pawlak, Z., Rough sets, Int. jour. of Info. And Comp. Sc. II,
(1982), 341-356.
[7] Pawlak, Z., Rough sets, Theoretical aspects of reasoning
about data, Kluwer academic publishers, London 1991.
[8] T Y Lin., Neighborhood System; A qualitative Theory for
fuzzy and Rough Sets, Advances in nachine Intelligence and
Soft Computing, Vol 4, Ed. Paul Wang, 1997, Duke University,
132-155.
[9] Tripathy B K & Mitra A, Some topological properties of
rough sets and their Application, Int. Jour. Granular Compt.,
Rough Sets and Intelligent Systems, Vol. 1, No. 4, 2010, Pp.
355-369.
[10] Wasserman, S. and K. Faust (1994). Social Network Analysis:
Methods and Applications. Cambridge: Cambridge
University Press.
40
41

You might also like