Abstract— The first place where the cheaters

Abstract— The first place where the cheaters find
a way to defame popular people is the social media. If a cheater gets the
complete access to others social network account, he can do whatever he
intended to do. This is because of the existing method of validation involved
in social media.  Normally these security
issues can be solved by following this proposed approach which was framed using
finite automata. At the end of this, the proposed approach called Automata
Based Fake Profile Identifier(ABFPR) is compared with the existing validation
methods and has given better results. This could be applied in the social media.

 

Keywords—Fake profile
Identifier; security of online social networks; finite automata; spam recognizer

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

 

I. INTRODUCTION

   There are Lots of researches are going on in
the field of data sharing and communication. Development in these streams laid
to the boom in the social networking sites. It is always good if the social
networks are used for good purpose. But it gives a bigger negative impact when
misused by unethical people. 123

 

In social media
like Face book, the main way how a fake person gets control over the genuine
user is by becoming friend with few of the victim’s actual friends in social
media 45. Common man and other relatives and friends will trust this
account as real seeing the list of friends. This makes the morphed account look
as real. Having added few real friends of the victim in friend list, the fake
account person further sends friend request to others and enlarges the friend
circle. Hence the fake account creator makes people to trust the fake profile
as real. This problem could be avoided if there is a process to verify if a
profile is valid or not before accepting

 

 

the friend requests.
In this paper, a new technique       based
on Finite Automata is proposed.

 

    We have proposed a Fake Profile
Identifier algorithm using Deterministic Finite Automata (DFA). We believe
that morphing, cheaters may mislead a human but, not the machine.  The algorithm proposed is based on the fact
that, an account pattern can be generated for each account based on the friend
list 6. This account pattern can be used to separate a fake friend from a
real friend.

 

 

II ACCOUNT PATTERN

    Let’s
see how an account pattern is generated

     Here are the points for account pattern
generation

 

  1. The country of user is the first thing
based on      which we generate a part of
the pattern

  2. The place where he completed the high
school education

  3. The place where he completed the middle
school

  4. The place he completed the higher
education if any

  5. Based on the currently having friends

 

A regular
expression is created for each account using the above five details.

   

This is the regular
expression will be the input for the fake account detector system.
Hence, every user will be having a deterministic finite automata system which
will get the request from the person and perform a account check and produce
the result that the requested friend is a true person or not. It will also
display the percentage of genuineness of that requested person. With the help
of this, lakhs of fake profiles could be identified and the users will feel
free to use the social networking accounts

 

III ACCOUNT PATTERN AS REGULAR EXPRESSION

 

1. His country is the first thing based on which we generate a part of
the pattern

              

Lets
consider he is from India consider India code is “IN” so if the requested
person is also from the same country certain genuine percentage will be added
this country will be very useful for rejecting the profile because only very
few people are friends with the other country people
 

 
 
 
 
 

Fig 1: 
DFA for country
 
This deterministic finite automata will
only allow if you are from India else this will go to the non-finite state
 
2. The
place where he completed the school education
Let’s
consider he is from Hindustan International school consider its code is HIN
so if the requested person  is also from
the same school persons genuineness increases

 
 
 
 

                               Fig 2 : DFA
for school
 
This
deterministic finite automaton will allow the transition from current state
to final state only if you are from Hindustan international school (HIN) else
this will go to the non-finite state. Even if a user creates his fake account
and identifies himself as an alumnus of Hindustan International School, his
ID will not be in the alumni association of the school. This will be
verified.
 
3. The place where he completed the
middle school
This
point is similar to the second point. This enables the users who study in
different schoold for higher classes to add the information for further
identification. 

 
 
 
 
 

                      Fig 3 : DFA for middle
school
If HMS stands for Hindustan Middle School, This
deterministic finite automata will only allow if you are from Hindustan
Middle School else this will go to the non-finite state
 
4. The place he completed the higher
education if any
Let’s
consider he is from Hindustan Institute Of Technology and Science represented
ad “HITS”. 

 
 
 
 
 
 

                                Fig 4 : DFA or higher education
This
deterministic finite automata will only allow if you are from HITS else this
will go to the non-finite state/
 
5. Based
on the currently having friends
 
   Initially there may not be any friends.
Based on the common information from the friend list pattern is generated.
This pattern is used as the key for Fake account detector system. Consider
the current friend pattern is prepared as “10(0+1)* + (0+1)* 0101” form
accessing his friend list this will also used to compute the same friends in
both the parties. The regular expression mentioned can be inferred as there
are few information fixed. Ie., The new friends can be either from the
pattern having all expressions starting with 10 or having all expressions
ending with 0101.This two information might be representing the college of
study or work place etc.,
 
 

6. Forming the regular expression

 

Now mix up all
the information and formulate a final regular expression and compute the
percentage of genuineness of the account. The old pattern Vs New pattern is shown in Figure 1.

In the above example, the regular
expression will be generated as

  

 

 

 

 

 

 

 

 

 

“IN-HIN-HMS-HITS-10(0+1)*
+ (0+1)* 0101”

This will be passed through the DFA of the
user for testing the profile genuineness finally give the result

Let’s take a look of the previous and the
current model and see what we have updated and how we will get the good fake
account recognizer as a result. The verification process in shown in Figure
2.

 

 

 

 

 

 

 

 

 

              

 

 

 

 

 

  

 

 

 

Figure 5.a Old Pattern

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 5.b New Pattern

 

The verification shown in
figure 2 has the DFA of IN-HIN-HMS-HITS-10(0+1)* +
(0+1)* 0101

Then consider
the BOB 1 ACCOUNT PATTREN  IN-HIN-HHMS-HIN -HITS -10001

BOB 2 ACCOUNT
PATTREN is IN-HIN-HMS-HITS-10110

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 6: Verification Process by ABFPR

 

 

Here BOB 1’s
account pattern doesn’t match with the required DFA of the genuine user and the
BOB 2’s account pattern matched so among these two persons BOB 1 is considered
as the fake profile and BOB 2 is the genuine profile. This is the power of the
ABFPR which will clearly differentiate the fake and the genuine profile

 

The Pseudo code
is as follows.

 

INPUT: A1,A2,A3………..An

PROCEDURE  fake profile detector

Get 
the account pattern

The make it into pieces like x-x1-x2-x3-friend pattern

If x is passed increment the genuine % by
10

If x1 is passed increment the genuine % by
10

If x2 is passed increment the genuine % by
10

If x3 is passed increment the genuine % by
10

If friend pattern  is passed increment the genuine % by 60

Then judge the genuine profile by
considering the genuine %

OUTPUT :-GENUINE ACCOUNTS A1,A3,……….

OUTPUT :-FAKE PROFILES A2, A4, A5,………….

 

 

IV
RESULTS

The proposed algorithm was tested with generated
test cases. It was checked with no of inputs varying from 10 users to 100
users. The results showed that the performance was better as the number of jobs
increased. The Figure 3 shows the performance by ABFPR.

 

 

                      Fig 7 : Graphical results
of accuracy

 

 

V CONCLUSION AND FUTURE WORK

The algorithm gave a better result
when compared with the existing manual verification techniques. But the regular
expression generated was quiet longer for the person who has friends with more
communities. Which could be worked on by using any algorithm may be by using
the period of study or period of work of any person

 

VI REFERENCES

1.       J A.
Malhotra, L. Totti, W. Meira Jr., P. Kumaraguru, and V. Almeida, “Studying User
Footprints in Different Online Social Networks,” The IEEE/ACM International Conference on
Advances in Social Networks Analysis and Mining.” 2012

2.       A.
R. Ahmad Rawashdeh, “Similarity Measure
for Social Networks – A
Brief Survey,” in CEUR Workshop Proceedings, At
Greensboro,NC,USA, 2015, vol. 1353.

3.       Bhume
Bhumiratana.”A Model for Automating Persistent Identity Clone in Online Social
Network,” 681–86. IEEE, 2011. doi:10.1109/TrustCom.2011.87.

4.       C.
G. Akcora, B. Carminati, and E. Ferrari, “Network and profile based measures
for user similarities on social networks,” in 2011 IEEE
International Conference on Information Reuse and Integration (IRI), 2011, pp.
292–298.

5.       Conti,
Mauro, Radha Poovendran, and Marco
Secchiero. “FakeBook: Detecting Fake Profiles in On-Line
Social Networks.” in Proceedings of the 2012 International Conference on
Advances in Social Networks Analysis and Mining (ASONAM 2012), 1071–1078. ASONAM ’12. Washington, DC, USA: IEEE Computer Society, 2012. doi:10.1109/ASONAM.2012.185.

6.       Mohamed
Torky1 , Ali Meligy2 , Hani Ibrahim3, Recognizing Fake Identities in Online
Social Networks Based on a Finite Automaton Approach, 12th International
Computer Engineering Conference (ICENCO), IEEE, 28-29 December, 2016, Cairo,
Egypt

Go Top
x

Hi!
I'm Rita!

Would you like to get a custom essay? How about receiving a customized one?

Check it out