Fast and Appropriate Social Network Analysis (SNA) tools ,techniques, are required to collect and classify
opinion scores on social networksites , as a grouping on wrong opinion may create problems for a society or
country . Social Network Analysis (SNA) is popular means for researcher as the number of users and groups
increasing day by day on that social sites , and a large group may influence other.In this paper, we
recommendhybrid model of opinion recommendation systems, for single user and for collective community
respectively, formed on social liking and influence network theory. By collecting thedata of user social networks
and preferenceslike, we designed aimproved hybrid prototype to imitate the social influence by like and sharing
the information among groups.The significance of this paper to analyze the suitability of ANN and Fuzzy sets
method in a hybrid manner for social web sites classifications, First, we intend to use Artificial Neural
Network(ANN)techniques in social media data classification by using some contemporary methods different
than the conventional methods of statistics and data analysis, in next we want to propagate the fuzzy approach
as a way to overcome the uncertainity that is always present in social media analysis . We give a brief overview
of the main ideas and recent results of social networks analysis , and we point to relationships between the two
social network analysis and classification approaches .This researchsuggests a hybrid classification model build
on fuzzy and artificial neural network (HFANN). Information Gain and three popular social sites are used to
collect information depicting features that are then used to train and test the proposed methods . This neoteric
approach combines the advantages of ANN and Fuzzy sets in classification accuracy with utilizing social data
and knowledge base available in the hate lexicons.
Gurgaon âĄď¸9711147426â¨Call In girls Gurgaon Sector 51 escort service
Â
Fuzzy AndANN Based Mining Approach Testing For Social Network Analysis
1. Vinay Viradia.et.al. Int. Journal of Engineering Research and App-lication www.ijera.com
ISSN : 2248-9622, Vol. 6, Issue 3, ( Part -5) March 2016, pp.48-52
www.ijera.com 48|P a g e
Fuzzy AndANN Based Mining Approach Testing For Social
Network Analysis
Vinay Viradia, NishidhChavda, Nilesh Dubey
Assistant Professor,Charotar University of Science &Technology,Changa . Gujarat. India
ABSTRACT
Fast and Appropriate Social Network Analysis (SNA) tools ,techniques, are required to collect and classify
opinion scores on social networksites , as a grouping on wrong opinion may create problems for a society or
country . Social Network Analysis (SNA) is popular means for researcher as the number of users and groups
increasing day by day on that social sites , and a large group may influence other.In this paper, we
recommendhybrid model of opinion recommendation systems, for single user and for collective community
respectively, formed on social liking and influence network theory. By collecting thedata of user social networks
and preferenceslike, we designed aimproved hybrid prototype to imitate the social influence by like and sharing
the information among groups.The significance of this paper to analyze the suitability of ANN and Fuzzy sets
method in a hybrid manner for social web sites classifications, First, we intend to use Artificial Neural
Network(ANN)techniques in social media data classification by using some contemporary methods different
than the conventional methods of statistics and data analysis, in next we want to propagate the fuzzy approach
as a way to overcome the uncertainity that is always present in social media analysis . We give a brief overview
of the main ideas and recent results of social networks analysis , and we point to relationships between the two
social network analysis and classification approaches .This researchsuggests a hybrid classification model build
on fuzzy and artificial neural network (HFANN). Information Gain and three popular social sites are used to
collect information depicting features that are then used to train and test the proposed methods . This neoteric
approach combines the advantages of ANN and Fuzzy sets in classification accuracy with utilizing social data
and knowledge base available in the hate lexicons.
Keywords-HFANN ,SNA ,SNS ,OWA , FRPR, Webmining , MATLAB
I. INTRODUCTION
Use of Social networking websites is popular
reven before the advent of websites like Facebook,
Twitter or Whatsapp. In [33] aextensive advent of
modeling social network data as undirected graphs
have been suggested and has been popularly
recognized. During the past eratremendousanalysis
has been executed on e.g. inter bonding of cluster
of members in social graphs [36] or distribution of
social networks [11].
However, the social weblog data analysis
of the social media sites is an upcoming field now ,
there exists numerous interesting World Wide
Webâs phenomenon, most of them are recorded by
BarabĂĄsi in his book âLinked: The New Science of
Networksâ. According to this book, âNetworks are
everywhere. Knowledge of them is mandatoryâ
[12].
Commonly social network analysis,
identical to the ones elaborated above, takes the
links between its actors as binary (1 if present, or 0
if not). Practically, not all the actors are associated
with same degree. For instance, hyperlinks between
the two websites belonging to the same
organization will demonstrate strong ties while
these same websites will build weak ties with the
third website belonging to some other organization.
Hence the cohesion between the
hyperlinks is different with different organization.
In traditional approach the social links between the
different organizations was given equal weight age.
MATLAB is a good tool to test non-
linear, multi-dimensional, correlated , widely
classified based data mining algorithm based on
ANN Principles ,for the testing of weblog based
mining problems related to social media .
It is generally accepted that - When a researcher
designed some algorithm that will generate test
cases to solve the problem through the MATLAB ,
it is possible to train the agents of ANN based
model to compare the results . This reduces the
efforts of mathematical testing for web mining.
As this study will compare theresultsets
from the ANN and Fuzzy set model , the working
principle is to use opinion and weights of the
groups in the social network pages , and it will
help to understand the uses of proposed hybrid
approach from researcher side .
Currently there is are good classification
approaches for Social Network Analysis (SNA)
as a combination of ANN and Fuzzy sets for social
network analysis . This paper will test this
approaches through MATLAB testing .
RESEARCH ARTICLE OPEN ACCESS
2. Vinay Viradia.et.al. Int. Journal of Engineering Research and App-lication www.ijera.com
ISSN : 2248-9622, Vol. 6, Issue 3, ( Part -5) March 2016, pp.48-52
www.ijera.com 49|P a g e
Early Work
Social networking sites are one that allow
users to create their own personal information page
containing information about self (real or virtual),
to communicate with other members of the same
website[4]. Communication is established in the the
form of electronic mail, comments written on each
othersâ personal pages, blog or pictures, or chatting
and messaging. Among the ten most popular social
sites Facebook and MySpace ranks the best web
sites in the world. Social Sites are extensively used
in many countries and include Orkut (Brazil),
Cyworld (Korea), and Mixi (Japan). The growth of
SNSs is influenced by the younger generation, with
Facebook originating as a college site [4] and
MySpace in the beginning only has 21 members in
early 2008 [48]. Nowadays the social sites are not
limited to the younger generation but also an
increasing number of aged people are actively
participating. The key motivating factor for using
Social sites is amiability, however, this implicate
that some types of anti social people may never be
interested in social sites[49]. Moreover, it seems
that sociability is helpful in such social sites [42]
and that female MySpace users have strong
inclination to male users and vice versa[40], which
recommend the basis for their effective
communication.
Fuzzy theory not only applicable to the
real life problems solutions but also can be applied
to all the social analysis. As argued by Brunelli and
Fedrizzi that most social analyses tools represent
adjacency relations in bidirectional(binary) form,
and presented âA fuzzy approach to social network
analysisâ (Brunelli&Fedrizzi, 2009). They
mathematically derived the fuzzy logic dimension
to demonstrate the associations in the social
network sites. They comprehend their fuzzy model
to represent the ordered weighted averaging
operators such as mean on m-ary fuzzy adjacency
relations.
Social Network Analysis combines the
concept of the sociogram (a pictorial arrangement
of associations in a social group) with basic
fundamentals of graph theory to inspect patterns of
correlation among people in different kinds of
networks, permitting quantitative resemblance
between different network structures.[12] There
exist a large number of research literature
explaining the use of Social Network Analysis.
Most of this literature work upon the basic
fundamentals of Social Network Analysis, that is
the development of abstract models of network
organization and the mathematical derivation of
quantitative measures of network
characteristics.[13] More
recentlytheworkwillexamine the association of
these quantitative compute with organizational
performance output.
Cummings and Cross, for example, come
out with the conclusion that that degree of
hierarchy, core-periphery structure, and structural
holes of leaders are associated negatively with
performance in 182 work groups in a large telecom
sector company,[14] and Aydin et al found that
extended network communication density was
related with maximal use of an electronic medical
record system by nurse practitioners and
physicianâs assistants.[15] There have also been
research showing how network parameters change
with due course of time. Shah, for example
demonstrated that network centrality decreased
after decreasing in a consumer electronics
firm,[16] whereas Burkhardt and Brass
documented growth in network centrality after
introduction of a new computer system in a federal
agency.[17]
II. RESEARCH METHODOLOGY
Social Netork is a social structure comprises of
individualsorgroups and users also know as
working or acting nodes , which are attached by
one or more assignedgroups based on like and their
strong relationships,such as friends, relatives
,office mates or college mates by their, knowledge
or prestige [1].
The data for the study and research
wereassimilated as part of a formal groups
selection from social media sites to evaluate the
opinion pool on some common issue inside or
outside a group , we took data from Facebook ,
Twitter users by downloading their web logs using
Informatica and other mining tools . Trained ANN
Agents used to assimilate data for two week in each
test case . They directly observed associations
among the users .groups based on
like,dislike,comments and their chatting patterns.
The methods of ANN based algorithms are tested
and than compared with Fuzzy sets
assumptionsbasedon multi-agent system theory of
ANN for social network analysis . Two well-
known and much used social networks Facebook
and Twitter users are analysed for this study .
A social network analysis user opinion
group decision making again tested by the
researcher using by advanced fuzzy reciprocal
preference relation (AFRPR) . The main novelty of
this method is that it can evaluate the importance
degree on some common issue by combining like
and comment scores . To do that, we used
correlated fuzzy and ANN sets to represent a
hybrid model for linking of relationship between
users and other members on social media .
In order to compute total sum of the
individual preference for like or comments on some
3. Vinay Viradia.et.al. Int. Journal of Engineering Research and App-lication www.ijera.com
ISSN : 2248-9622, Vol. 6, Issue 3, ( Part -5) March 2016, pp.48-52
www.ijera.com 50|P a g e
common issue, it is necessary to determine the
scores associated to each user and group on social
media . A general assumption used by trained
agents of MATLAB is that the weights of users or
groups is usually changed by other users like and
dislikes . However, this assumption may not be
wrong in real situation, and therefore it is required
to check it again using Fuzzy sets. Trust and
friendship is a key element for groups for like and
comments , and as we propose a hybrid SNA
methodology for the representation and testing the
relationships between users in a social group so
We tested it by Multi-Layer Perceptions method in
MATLAB based on principles of ANN . The input
scores of thesemulti agent networks is operatedby
the next layers and the output value checked by the
next layer . For similar types of network, the
information is relayed forward from many layers,
respectively. The data from the input layer is
relayed to the hidden layer and output layer
operated this data that is transferred to it in the
activation function and processed the final score .
The acquired output is measured with the desired
output in both the cases . This process is repeated
for 200 test cases .
III. RESULTS
Results of proposed hybrid approach by
computing overall frequency of user or groups on
like or dislike recorded . For example, as the
individual,user1like theuser2 comments and share
it for next 20groups or users and the user2likes
user1 in reply , the user5 and user6 never
comments on the user2 opinion. As they never
visited his profile ,that proved our test case ,
because the opinion score of the individuals are
different from each other.
Observed
User
Like Centrality Sharing Centrality
1 0.0924984 0.5934500
2 0.1096495 0.5937475
3 0.0587399 0.5428589
4 0.0408626 0.5428571
5 0.0285575 0.5135135
6 0.0485450 0.4418605
7 0.0871415 0.5135135
8 0.0579342 0.5277778
9 0.0428061 0.5757576
10 0.0489116 0.4871795
11 0.0631892 0.6333555
12 0.0421436 0.5277778
13 0.0419247 0.5428688
14 0.0645526 0.5937601
15 0.1034649 0.6333555
16 0.03970020.5588198
17 0.0834470 0.5937500
18 0.0385025 0.5588235
19 0.05251900.5588230
Table 1
When we look at the like testing in the
Table 1, we see that the user2 and user15 have high
ranklike grade similar to others with the values
0.1096495 and 0.1034649, respectively. It
iselucidated that user2 and user15, act as the leader
intheir social network group . It is interpreted that
the individuals groups or users , who has the higher
like correlations , are the highly active users in
their groups and they more influential also . When
we look at the sharing testing ofusers the user11
and user15 have the highest closeness degree
(0.6333555) and retrievingthe information by those
user who are having higher degree than other users
or groups . This demonstrated that the links . 11
and 15 are the most dominant loops. The loop,
which the sharing centralization is lower
(0.4418605), is the user6, and it interpret that this
user6, is the most inactive link in this social
network group.
4.1 TEST1 For Opinion differences (Using
ANN)
QAP correlation Value SignifAbg SD
D(large) D(small) Nperm
Pearson Correlation . 493 .000 .002 .059
.000 1.000 2485
4.2 TEST2 For Opinion differences (Using
Fuzzy Sets)
QAP correlation Value SignifAbg
SD D(large) D(small) Nperm
Pearson Correlation -.133 .000 -.002
. 047 .989 .000 2494
IV. CONCLUSIONS
Fast and hybrid social network analysis
techniques are needed to mine opinion scores on
social networks as a grouping on wrong opinion
may create problems for a society or country .
Social Network Analysis (SNA) can be used as an
important tool for researchers , as the number of
users and groups increasing day by day on that
social sites , and a large group may influence
others, but the necessary information is often
distributed and hidden on social site servers , so
there is a need to design some new approaches for
collection and analysis the social web data . The
paper hypotheses are tested on Fuzzy set and ANN
4. Vinay Viradia.et.al. Int. Journal of Engineering Research and App-lication www.ijera.com
ISSN : 2248-9622, Vol. 6, Issue 3, ( Part -5) March 2016, pp.48-52
www.ijera.com 51|P a g e
based models to improve finding relation and to
prove that there is meaningful and good
correlation between the like and share of social
media users on a common issue while chatting or
commenting , even like and dislike ratio is very
much similar and mean of final inputs we used for
output layer. Everytestcase results is significant
and there are higher degree of associations. This
implies that the methods of Artificial Neural
Networks, may be applied relatively, Correlation
between the Artificial Neural Networksresults and
Fuzzy set results may also be elucidated in form of
Artificial Neural Networks . Association of both
the methods improved the result upto4.5% , as user
data, transferred to output layer in ANN with the
connection weights will permit for interpreting the
structure in the hidden layer of Artificial Neural
Networks and also permit to make input for Fuzzy
set layer . Furthermore to this, when the internal
dynamics of Artificial Neural Networksand Fuzzy
are analyzed, it maygenerate new methods named
hybrid classification model using fuzzy and
artificial neural network (HFANN) by us for this
study . In the future studies, more efficient results
will be derived as we will increase the amount of
web log data taken from social network sites .For
future work, we extendthis method can be applied
to various other social networking websites.
REFERENCES
[1]. Meishan, H.; Aixin, S.; EeâPeng, L.
Commentsâ Oriented Document
Summarization: Understanding
Documents with Readersâ Feedback.
SIGIRâ08, 2008, 291â298
[2]. Personalization: Collaborative Filtering
vs. Prediction Based on Benefit Theory.
Available:
http://myshoppal.typepad.com/blog/2007/
11/persona lization.html. [3] Andriy, S.;
Jonathan, G.; Bamshad, M.; Robin, B.
Personalized recommendation in social
tagging systems using hierarchical
clustering. In Proceedings of the 2008
ACM conference on Recommender
systems, 2008, 259â266.
[3]. Boyd, D., Ellison, N.: Social network
sites: definition, history and scholarship. J.
Comput. Mediat. Commun. 13(1), 210â
230 (2007)
[4]. Cauvery, N., Viswanatha, K.V.: Routing
in dynamic network using ants genetic
algorithm. IJCSNS Int. J. Comput. Sci.
Netw. Secur. 9(3), 194â200 (2009)
[5]. Choudhuri, T., Pentland, A.: Sensing and
modelling human networks using the
sociometer. In: Proceedings of 7th IEEE
Symposium on Wearable Computing,
New York (2003)
[6]. Czinkota, M.R., Knight, G.A., Liesch,
P.W., Steen, J.: Positioning terrorism in
management and marketing: research
propositions. J. Int. Manag. 11(4), 581â
604 (2005)
[7]. De Nooy, W., AMrvar, A., Batagelig, V.:
Explorating SNA with pajek. Cambridge
University Press, Cambridge/New York
(2004)
[8]. Deng, Y., Tong, H.: Dynamic shortest
path algorithm in stochastic traffic
networks using PSO based on fluid neural
network. J. Intell. Learn. Syst. Appl. 3,
11â16 (2011)
[9]. M. ÄiriÄ, J. IgnjatoviÄ, S. BogdanoviÄ,
Uniform fuzzy relations and fuzzy
functions, Fuzzy Sets and Systems 160
(2009) 1054â1081.
[10]. M. ÄiriÄ, A. StamenkoviÄ, J. IgnjatoviÄ, T.
PetkoviÄ, Fuzzy relation equations and
reduction of fuzzy automata, Journal of
Computer and System Sciences (2009),
doi: 10.1016/j.jcss.2009. 10.015.
[11]. M. ÄiriÄ, J. IgnjatoviÄ, S. BogdanoviÄ,
Uniform fuzzy relations and fuzzy
functions, Fuzzy Sets and Systems 160
(2009) 1054â1081.
[12]. M. ÄiriÄ, A. StamenkoviÄ, J. IgnjatoviÄ, T.
PetkoviÄ, Fuzzy relation equations and
reduction of fuzzy automata, Journal of
Computer and System Sciences (2009),
doi: 10.1016/j.jcss.2009. 10.015.
[13]. Cummings JN, Cross R. Structural
properties of work groups and their
consequences for performance. Soc
Networks. 2003;25:197â210.
[14]. Aydin CE, Anderson JG, Rosen PN,
Felitti VJ, Weng HC. Computers in the
consulting room: a case study of clinician
and patient perspectives. Health Care
Manag Sci. 1998;1:61â74. [PubMed]
[15]. Shah PP. Network destruction: the
structural implications of
downsizing. Acad Management J.
2000;43:101â112.
[16]. Burkhardt ME, Brass DJ. Changing
patterns or patterns of change: the effects
of a change in technology on social
network structure and
power. Administrative Sci Quarterly.
1990;35:104â127.
[17]. Ucinet for Windows: Software for Social
Network Analysis. Cambridge, Mass:
Analytic Technologies; 2002.
[18]. Krackhardt D, Blythe J, McGrath C.
KrackPlot 3.0: an improved network
5. Vinay Viradia.et.al. Int. Journal of Engineering Research and App-lication www.ijera.com
ISSN : 2248-9622, Vol. 6, Issue 3, ( Part -5) March 2016, pp.48-52
www.ijera.com 52|P a g e
drawing program. Connections.
1994;17:53â55.
[19]. Watts DJ. Small Worlds: The Dynamics
of Networks Between Order and
Randomness. Princeton, NJ: Princeton
University Press; 1999.
[20]. G. B. Davis, K. M. Carley, Clearing the
FOG: Fuzzy, overlapping groups for
social networks, Social Networks 30
(2008) 201â212. 13. M. G. Everett, S. P.
Borgatti, Regular equivalences: general
theory, Journal of Mathematical Sociology
18 (1994) 29â52. 14. T. F. Fan, C. J. Liau,
T. Y. Lin, Positional analysis in fuzzy
social networks, in: Proceedings of the
Third IEEE International Conference on
Granular Computing, 2007, pp. 423â428.
15. T. F. Fan, C. J. 23Liau, T. S. Lin, A
theoretical investigation of regular
equivalences for fuzzy graphs,
International Journal of Approximate
Reasoning 49 (2008), 678â688.
[21]. BarabĂĄsi, A.-L. (2003). Linked: The New
Science of Networks . Perseus Publishing.
[22]. Brunelli, M., &Fedrizzi, M. (2009). A
Fuzzy Approach to Social Network
Analysis. 2009 International Conference
on Advances in Social Network Analysis
and Mining, (pp. 225-230).
[23]. Freeman, L. C. (2004). The Development
Of Social Network Analysis: A Study in
the Sociology of Science. BookSurge,
LLC.
[24]. Gonzalez-Bailon, S. (2009). Opening the
black box of link formation: Social factors
underlying the structure of the web. Social
Networks, 31.4, 271â280.
[25]. Noah E. Friedkin and Eugene C. Johnsen,
âSocial Influence Networks and Opinion
Changeâ, Advances in Group Processes,
vol. 16, pp.1-29, 1999
[26]. Pan Hui and Sonja Buchegger,
âGroupthink and Peer Pressure: Social
Influence in Online Social Network
Groupsâ, In Proceeding International
Conference on Advances in Social
Networks Analysis and Mining
(ASONAM), Athens, Greece, July, 2009.
[27]. David Kempe, Jon Kleinberg and Eva
Tardos, âMaximizing the Spread of
Influence through a Social Networkâ, In
Proc. 9th ACM SIGKDD Intl. Conf. on
Knowledge Discovery and Data Mining,
2003.
[28]. Jon Kleinberg, âCascading Behavior in
Networks: Algorithmic and Economic
Issuesâ, In Algorithmic Game Theory (N.
Nisan, T. Roughgarden, E. Tardos, V.
Vazirani, eds.), Cambridge University
Press, 2007. [32] David Easley and Jon
Kleinberg, âNetworks, Crowds, and
Markets: Reasoning About a Highly
Connected Worldâ, Cambridge University
Press, pages 497-517, 2010.
[29]. Duncan J. Watts and Steven Strogatz,
âCollective dynamics of âsmallworldâ
networksâ, Nature 393 (6684): 40910,
1998.
[30]. Watts, D.J., Strogatz, S.H.: Collective
dynamics of âsmall-worldâ networks.
Nature 393(6684), 440â442 (1998)
[31]. Wendling, F., Ansari-Asl, K., Bartolomei,
F., Senhadji, L.: From EEG signals to
brain connectivity: A model-based
evaluation of interdependence measures. J
Neurosci Methods 183(1), 9â18 (2009)
[32]. White, D.R., Harary, F.: The cohesiveness
of blocks in social networks: Node
connectivity and conditional density.
SociolMethodol 31(1), 305â359 (2001)
[33]. WĂźst, S., Kasten, E., Sabel, B.A.:
Blindsight after optic nerve injury
indicates functionality of spared fibers. J
CognNeurosci 14(2), 243â253 (2002)