Detection of Fake Profiles On Social Networks: Nadir Mahammed April 25, 2024

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

Detection of Fake Profiles on Social Networks

Nadir MAHAMMED
April 25, 2024

Contents
1 Context and Problem 2
1.1 Global Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Local Context (Algeria) . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2 Key Concepts 3
2.1 Types of Fake Profiles . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Machine Learning and Metaheuristics . . . . . . . . . . . . . . . 3

3 Our Contributions 4
3.1 First Contribution: BIA/ML Hybridization . . . . . . . . . . . . 4
3.2 Second Contribution: NLP/BIA+ML Approach . . . . . . . . . . 4
3.3 Third Contribution: BIA-Based Detection . . . . . . . . . . . . . 4
3.4 Fourth Contribution: Australian Firehawk Optimization . . . . . 4
3.5 Fifth Contribution: Grey Wolf Optimization . . . . . . . . . . . . 4

4 Work Ahead 5

5 References 6

1
1 Context and Problem
This chapter provides an overview of the global and local context of internet
usage and the rise of social media. It highlights the significant challenges posed
by fake profiles on online social networks, focusing on their impact and the
necessity for efficient detection mechanisms.

1.1 Global Context


The chapter discusses the growth of connected devices and services worldwide,
reasons for internet use, and the types of commonly used websites and apps.

1.2 Local Context (Algeria)


The chapter examines internet usage and connected services in Algeria, provid-
ing a regional perspective on the broader global trends.

1.3 Problem Statement


The main problem addressed in this work is the detection of fake profiles on
social networks, emphasizing the need for effective detection methods.

2
2 Key Concepts
This chapter introduces key concepts related to profiles on social networks and
fake profiles. It covers the evolution of social network profiles, various types of
fake profiles, and their purposes.

2.1 Types of Fake Profiles


- **Suck Puppets**: Accounts created to deceive users. - **Honey Profiles**:
Attractive profiles used to lure users. - **Bots**: Automated accounts per-
forming tasks at scale.

2.2 Machine Learning and Metaheuristics


The chapter also provides a foundation in machine learning, multi-objective op-
timization, and metaheuristic algorithms, essential for understanding the pro-
posed detection methods.

3
3 Our Contributions
The core contributions of this work are detailed across multiple sections, each
introducing a novel approach to fake profile detection.

3.1 First Contribution: BIA/ML Hybridization


- **Proposition**: A hybrid system combining bio-inspired algorithms (BIA)
and machine learning (ML) for fake profile detection. - **System Flowchart**:
An illustrative flowchart of the proposed system. - **Satin Bowerbird Opti-
mizer**: Introduction and pseudocode for the Satin Bowerbird Optimizer used
in the system. - **Experimentation**: Results from various machine learning
classifiers (e.g., SVM, Naive Bayes) applied to the Facebook dataset.

3.2 Second Contribution: NLP/BIA+ML Approach


- **Proposition**: Integration of Natural Language Processing (NLP) with BIA
and ML for enhanced detection. - **System Flowchart**: Flowchart depicting
the proposed NLP-enhanced system. - **Particle Swarm Optimization**: De-
tailed pseudocode and application. - **Experimentation**: Evaluation using
Twitter dataset and comparison of results.

3.3 Third Contribution: BIA-Based Detection


- **Proposition**: Purely bio-inspired algorithm approach. - **Tasmanian
Devil Optimization**: Introduction and pseudocode. - **Experimentation**:
Results from applying the algorithm to the Twitter dataset.

3.4 Fourth Contribution: Australian Firehawk Optimiza-


tion
- **Proposition**: Firehawk optimization algorithm for fake profile detection.
- **Firehawk Optimizer**: Pseudocode and natural-to-artificial transition. -
**Experimentation**: Results from applying the algorithm.

3.5 Fifth Contribution: Grey Wolf Optimization


- **Proposition**: Grey Wolf optimization algorithm. - **Grey Wolf Opti-
mizer**: Pseudocode and detailed comparison with other methods. - **Exper-
imentation**: Results and performance metrics.

4
4 Work Ahead
This chapter outlines future directions for research and development in the
detection of fake profiles. Potential areas include: - Integration of NLP and
sentiment analysis. - Detection of coordinated group activities. - Expanding to
fake news detection.

5
5 References
• Roy, P. K., & Chahar, S. (2020). Fake profile detection on social network-
ing websites: a comprehensive review. IEEE Transactions on Artificial
Intelligence, 1(3), 271-285.
• Wani, M. A., & Jabin, S. (2017). A sneak into the Devil’s Colony-Fake
Profiles in Online Social Networks. arXiv preprint arXiv:1705.09929.

• Talbi, E. G. (2021). Machine learning into metaheuristics: A survey and


taxonomy. ACM Computing Surveys (CSUR), 54(6), 1-32.
• And more...

You might also like