| Author: | Yao, Zhongyu |
| Title: | Fake follower detection : effectiveness of feature-based methods and potential of data augmentation approach |
| Advisors: | Chung, Fu-lai Korris (COMP) |
| Degree: | M.Sc. |
| Year: | 2020 |
| Subject: | Online social networks Social media Computer security Hong Kong Polytechnic University -- Dissertations |
| Department: | Department of Computing |
| Pages: | [68] pages : illustrations |
| Language: | English |
| Abstract: | Online Social Network (OSN) is undoubtedly the first choice for people to exchange information in modern society. People rely on these platforms for news, updates and more. Although social networks can provide very effective information, at the same time, they are rapidly becoming a place for disseminating misinformation, rumors, unsolicited news, etc. It is essential to detect and prevent these issues in advance to maintain a healthy social network ecosystem. To achieve this goal, there are three main steps, (1) a series of profiles of fake and legitimate accounts are collected in Twitter, (2) the features of these accounts records are analyzed and implemented in a series of algorithms of machine learning, and (3) The performances against different learning algorithms are assessed thoroughly. The algorithms includes Random forest, SVM, AdaBoost, Decision trees, Naive Bayes , K-NN and Neural Network. The best results of accuracy, precision, recall, F-M, and AUC are 0.988, 0.992, 0.99, 0.99, and 0.99 separately. Experimental results showed that features in these algorithms are effective. Meanwhile, I investigated the data augmentation methods for text, which are deep-learning-based and more concerned by the researchers recently. It is also helpful to confirm whether the account is friendly to the ecosystem. The experimental result has an accuracy of 0.9836 and a loss of 0.0486. |
| Rights: | All rights reserved |
| Access: | restricted access |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 5814.pdf | For All Users (off-campus access for PolyU Staff & Students only) | 1.62 MB | Adobe PDF | View/Open |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/11366