Scribd
Upload
95 views

Fake Profile Identification - Abstract

This document proposes a model using support vector machines to classify social media accounts as fake or genuine. It notes that Facebook reported removing over 500 million fake accounts in the first quarter of 2018 alone. The proposed model aims to automatically classify accounts at scale without manually evaluating each one. It extracts relevant features from accounts and feeds them into a regularly trained classifier that determines if an account is fake or not based on the training data. The model achieves over 90% accuracy on a large dataset. Future work could address the high variance issue by increasing the training data available.

Uploaded by

V.durga Revathi pavani
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
95 views

Fake Profile Identification - Abstract

This document proposes a model using support vector machines to classify social media accounts as fake or genuine. It notes that Facebook reported removing over 500 million fake accounts in the first quarter of 2018 alone. The proposed model aims to automatically classify accounts at scale without manually evaluating each one. It extracts relevant features from accounts and feeds them into a regularly trained classifier that determines if an account is fake or not based on the training data. The model achieves over 90% accuracy on a large dataset. Future work could address the high variance issue by increasing the training data available.

Uploaded by

V.durga Revathi pavani
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

FAKE PROFILE IDENTIFICATION

Abstract:
The social network, a crucial part of our life is plagued by online impersonation
and fake accounts. According to the ‘Community Standards Enforcement Report’ published
by Facebook on March 2018, about 583 million fake accounts were taken down just in
quarter 1 of 2018 and as many as 3-4% of its active accounts during this time were still fake.
In this project, we propose a model that could be used to classify an account as fake or
genuine. This model uses Support Vector Machine as a classification technique and can
process a large dataset of accounts at once, eliminating the need to evaluate each account
manually. The community of concern to us here is Fake Accounts and our problem can be
said to be a classification or a clustering problem.
Introduction:
In the present generation, the social life of everyone has become associated with the online
social networks. Adding new friends and keeping in contact with them and their updates has
become easier. The online social networks have impact on the science, education, grassroots
organizing, employment, business, etc. Researchers have been studying these online social
networks to see the impact they make on the people. Teachers can reach the students easily
through this making a friendly environment for the students to study, teachers nowadays are
getting themselves familiar to these sites bringing online classroom pages, giving homework,
making discussions, etc. which improves education a lot. The employers can use these social
networking sites to employ the people who are talented and interested in the work, their
background check can be done easily.

Existing System:
• Naive Bayes algorithm having less accuracy.

Disadvantages:
Because of Privacy Issues the Facebook dataset is very limited and a lot of details
are not made public.
Proposed System:
The Application Domain of the following project was Community Detection.
Community detection is key to understanding the structure of complex networks, and
ultimately extracting useful information from them
In this project, we came up with a framework through which we can detect a fake profile
using machine learning algorithms so that the social life of people become secured.
1. Classification starts from the selection of profile that needs to be classified.
2. Once the profile is selected, the useful features are extracted for the purpose of
classification.
3. The extracted features are then fed to trained classifier.
4. Classifier is trained regularly as new data is fed into the classifier.
5. Classifier then determines whether the profile is genuine or fake.
6. The result of classification algorithm is then verified and feedback is fed back into the
classifier.
7. As the number of training data increases the classifier becomes more and more accurate in
predicting the fake profiles.

Advantages:
• The social networking sites are making our social lives better but nevertheless there
are a lot of issues with using these social networking sites.
• The issues are privacy, online bullying, potential for misuse, trolling, etc. These are
done mostly by using fake profiles. In
• In this project, we came up with a framework through which we can detect a fake
profile using machine learning algorithms so that the social life of people become
secured.

System Architecture:

Select The
Profile To Be
tested

Train the classifier Extract The Attributes


using the feedback Required

Pass it through the


Get the feedback
trained classifier
of the result

Determine real/
fake

Software and Hardware Requirements:


Hardware:
• OS – Windows 7, 8 and 10 (32 and 64 bit)
• RAM – 4GB

Software:
• Python
• Anaconda

CONCLUSIONS AND FUTURE WORK


The model presented in this project demonstrates that Support Vector Machine (SVM) is an
elegant and robust method for binary classification in a large dataset. Regardless of the non-
linearity of the decision boundary, SVM is able to classify between fake and genuine profiles
with a reasonable degree of accuracy (>90%). This method can be extended on any platform
that needs binary classification to be deployed on public profiles for various purposes. This
project uses only publicly available information which makes it convenient for organizations
that want to avoid any breach of privacy, but organizations can also use private data available
to them to further extend the capabilities of the proposed model.
Future Work
Since we have limited data to train the classifier, our approach is facing a high variance
problem which can be observed in the learning curve as follows High variance problems can
usually be mitigated by increasing the size of the dataset which should not be of much
concern to Social Networks Organizations which already have fairly large datasets

You might also like