Chapter 1

INTRODUCTION

INTRODUCTION

• In 2017 Facebook reached a total population of 2.46 billion users making

it the most popular choice of social media [1]. Social media networks
make revenues from the data provided by users. The average user does
not know that their rights are given up the moment they use the social
media network's service. Social media companies have a lot to gain at
the expense of the user. Every time a user shares a new location, new
photos, likes, dislikes, and tag other users in content posted, Facebook
makes revenue via advertisements and data. More specifically, the
average American user generates about $26.76 per quarter [2]. That
number adds up quickly when millions of users are involved.

In today's digital age, the ever-increasing dependency on computer

technology has left the average citizen vulnerable to crimes such as data
breaches and possible identity theft. These attacks can occur without
notice and often without notification to the victims of a data breach. At this
time, there is little incentive for social networks to improve their data
security. These breaches often target social media networks such as
Facebook and Twitter. They can also target banks and other financial
institutions.

1
 Chapter 2
ABSTRACT

we use machine learning, namely an artificial neural network to determine

what are the chances that Facebook friend request is authentic or not. We

also outline the classes and libraries involved. Furthermore, we discuss

the sigmoid function and how the weights are determined and used.

Finally, we consider the parameters of the social network page which are

utmost important in the provided solution.

The other dangers of personal data being obtained for fraudulent purposes

is the presence of bots and fake profiles. Bots are programs that can

gather information about the user without the user even knowing. This

process is known as web scraping. What is worse, is that this action is

legal. Bots can be hidden or come in the form of a fake friend request on a

social network site to gain access to private information.

2
EXISTING SYSTEM :-

• Malicious users create fake profiles to phish login information from

unsuspecting users. A fake profile will send friend requests to many
users with public profiles. These counterfeit profiles bait unsuspecting
users with pictures of people that are considered attractive. Once the
user accepts the request, the owner of the phony profile will spam
friend requests to anyone this user is a friend.
• The fake profile's contents typically have links that lead to an external
website where the damage happens. An unaware curious user clicking
the bad link will damage their computer. The cost can be as simple as
catching a virus to as bad as installing a rootkit turning the computer
into a zombie. While Facebook has a rigorous screening to keep these
fake accounts out, it only takes one fake profile to damage the
computers of many.
PROPOSED SYSTEM :-
• In our solution, we use machine learning, namely an artificial neural
network to determine what are the chances that a friend request is
authentic or not.
• We utilize Microsoft Excel to store old and new fake data profiles. The
algorithm then stores the data in a data frame. This collection of data
will be divided into a training set and a testing set. We would need a
data set from the social media sites to train our model.
• For the training set, the features that we use to determine a fake profile
are Account age, Gender, User age, Link in the description, Number of
messages sent out, Number of friend requests sent out, Entered
location, Location by IP, Fake or Not. Each of these parameters is
tested and assigned a value. For example, for the gender parameter if
the profile can be determined to be a female or male a value of (1) is
assigned to the training set for Gender. The same process is applied to
other parameters. We also use the country of origin as a factor

3
• We then determine the Number of messages sent out parameter by
dividing the number of messages sent by the age of the account. We
then determine the Number of friend requests sent out parameter by
dividing the Number of friend computing and used primarily for multi-
dimensional matrix multiplication as we are dealing with a large amount
of numbers that are very dependent on each other.

ADVANTAGES OF PROPOSED SYSTEM:

• Vote Trust uses a voting based system that pulls user activities to find
fake profiles using trust-based vote assignment and global votes total.
It is considered as the first line of defense due to limitations which
include real accounts that were already compromised being sold

4
 SOFTWARE REQUIREMENT SPECIFICATION

User Requirements

Requirement Specification plays an important role to create quality software

solution. Requirements are refined and analysed to assess the clarity.
Requirements are represented in a manner that ultimately leads to successful
software implementation.
Software Requirements
The software requirements specification is produced at the end of the analysis
task. Software requirement is a difficult task, For developing the Application

1. Python
2. Django
3. Mysql
4. Mysqlclient
5. WampServer 2.4
• Technologies and Languages used to Develop
-- Python
Hardware Requirements
This is an project so hardware plays an important role. Selection of hardware also
plays an important role in existence and performance of any software. The size and
capacity are main requirements.

• Operating System supported by

1. Windows 7
2. Windows XP
3 . Windows 8
• Processor – Pentium IV or higher
• RAM -- 256 MB
• Space on Hard Disk -- Minimum 512 MB

5
Module Description:-

 Admin Module: Admin will login to application by using

username as ‘admin’ and password as ‘admin’ and then perform
below actions.
a) Generate ANN Train Model: Admin will upload profile dataset to
ANN algorithm to build train model. This train model can be used
to predict fake or genuine account by taking new account test data.
b) View ANN Train Dataset: Using this module admin can view all
dataset used to train ANN model.
 User Module: Any user can use this application and enter test data
of new account and call ANN algorithm. ANN algorithm will take
new test data and applied train model to predict whether given test
data contains fake or genuine details.

6
 Chapter 4

DESIGN

INTRODUCTION

Software design sites at the technical kernel of the software engineering process and
is applied regardless of the development paradigm and the area of application. Design
is the first step in the development phase for any engineered product or system. The
designer’s goal is to produce a model or representation of an entity that will later be
built. Beginning, once system requirements have been specified and analysed, system
design is the first of the three technical activities – design, code and test that is
required to build and verify software.

During design, progressive refinement of data structure, program structure and

procedural details are developed, reviewed and documented. System design can be
viewed from either technical or project management perspective. From the technical
point of view, design is comprised of four activities – architectural design, data
structure design, interface design, procedural design.

Feasibility Study
Feasibility study is accompanied once the difficult is obviously understood.
The feasibility study which is a great level lozenge version of the whole system
analysis and design procedure. The independent is to define whether the planned
system is possible or not and it benefits us to the least expense of how to resolve the
problem and to govern, if the Problem is wealth solving.

The following are the three important tests that have been conceded out for
feasibility Study.

 Technical Feasibility
 Economic feasibility
 Operational feasibility

7
TECHNICAL FEASIBILITY

In the technical feasibility study, one has to assess whether the implemented
system can be established using existing technology or not. It is intended to
implement the implemented system in JSP. The project enabled is theoretically
feasible since the following reasons.

 All needed technology exists to improve the system.

 The existing system is so malleable that it can be advanced further.

ECONOMIC FEASIBILITY

As a portion of this, the expenses and profits related with the implemented
systems are to be associated. The project is carefully feasible only if tangible and
intangible assistances balance the cost. We can say the implemented system is
feasible founded on the following grounds.

 The charge of developing the filled system is sensible.

 The cost of hardware and software for the application is less.
OPERATIONAL FEASIBILITY

This project is operationally feasible for there is necessary support from the
project organization and the users of the implemented system .Implemented system
absolutely does not damage and determination not create the corrupt results and no
problem will ascend after implementation of the system.

User-friendly

Customer will use the forms for their various transactions i.e. for adding new
routes, viewing the routes details. Also the Customer wants the reports to view the
various transactions based on the constraints. These forms and reports are generated
as user-friendly to the Client.

Reliability

The package wills pick-up current transactions on line.

Regarding the old transactions, User will enter them in to the system.

Security

8
The web server and database server should be protected from hacking, virus etc.

UML DIAGRAM’S

CLASS DIAGRAMS

4.3.2 USE CASE DIAGRAMS

9
4.3.3 SEQUANCE DIAGRAMS

Fig 4.3.3:sequence diagrams

10
 CHAPTER 5

SCREEN SHOTS & CONCLUSION & & REFERENCES

In above screen click on ‘ADMIN’ link to get below login screen

11
In above screen enter admin and admin as username and password to
login as admin. After login will get below screen

In above screen click on ‘Generate ANN Train Model’ to generate

training model on dataset. After clicking on that link you can see server
console to check ANN processing details with accuracy

12
In above black console we can see all ANN details.

In above screen we can see ANN got 98% accuracy to train all Facebook
profile. Now click on ‘View Ann Train Dataset’ link to view all dataset
details

13
In above screen we can see all train data and scroll down to view all
records. Now ANN train model is ready and you can logout and click on
‘User’ link to get below screen.

In above screen enter some test account details to get

prediction/identification from ANN. You can use below records to check

14
10, 1, 44, 0, 280, 1273, 0, 0
10, 0, 54, 0, 5237, 241, 0, 0
7, 0, 42, 1, 57, 631, 1, 1
7, 1, 56, 1, 66, 623, 1, 1

For above input will get below result

In above screen we can see the result predicted as genuine account

15
For above account details we got below result

In above screen we got result as fake for given account data

16
CONCLUSION

• we use machine learning, namely an artificial neural network to

determine what are the chances that a friend request is authentic are
or not. Each equation at each neuron (node) is put through a Sigmoid
function. We use a training data set by Facebook or other social
networks. This would allow the presented deep learning algorithm to
learn the patterns of bot behavior by back propagation, minimizing the
final cost function and adjusting each neuron's weight and bias.

• In this paper, we outline the classes and libraries involved. We also

discuss the sigmoid function and how are the weights determined and
used. We also consider the parameters of the social network page
which are the most important to our solution.

17
FUTURE SCOPE :

• Each input neuron would be a different, previously chosen feature of

each profile converted into a numerical value (e.g., gender as a binary
number, female 0 and male 1) and if needed, divided by an arbitrary
number (e.g., age is always divided by 100) to minimize one feature
having more influence on the result than the other. The neurons
represent nodes. Each node would be responsible for exactly one
decision-making process

18