|
|
| Line 1: |
Line 1: |
| − | = Project Team =
| + | Automated accounts on Twitter known as Twitterbots are being used for malicious purposes. There is a particular concern in the use of Twitterbots in information warfare to influence politics or cause civil unrest. Research has been conducted into methods that could be used to identify Twitterbots with the aim of removing their presence from Twitter. The aim of this project is to evaluate the Twitterbot identification methods that have been proposed and to propose a new framework for Twitterbot identification. The techniques examined include human-based methods, supervised machine learning, activity correlation, monitoring user activity and clustering. In addition a botnet case study has been examined that shows that a targeted approach to Twitterbot identification can be effective |
| − | David Hubczenko
| |
| − | | |
| − | = Supervisors =
| |
| − | Dr Matthew Sorell
| |
| − | | |
| − | = Introduction =
| |
| − | Social media is a widely-used platform to connect people all around the world. It allows people to share their ideas with a broad audience. Currently, one of the most popular social media platforms is Twitter. The main feature of Twitter is the ability to post short, 140-character limited messages called tweets.
| |
| − | | |
| − | On Twitter, automated accounts known as Twitterbots are being used for malicious purposes. One of the major concerns is the use of Twitterbots in information warfare to influence politics or cause civil unrest. This is achieved through social engineering attacks on Twitter. Twitterbots facilitate social engineering because they can produce content on a much larger scale than an ordinary account. Psychology research indicates that a large quantity of messages can highly increase the persuasive power of the messages [1]. Other malicious activities of Twitterbots include spreading spam and inflating the followers of Twitter accounts.
| |
| − | | |
| − | One recent example of malicious Twitterbot activity is in the 2016 United States presidential elections. It has been suggested that the elections were potentially influenced by Twitterbots [2]. It has been estimated that a third of the pro-Trump tweets and a fifth of the pro-Clinton tweets during the election were generated by Twitterbots [2]. Another example is the use of the Twitterbots for economic gain. In 2014 Twitterbots were used to make the company Cynk appear popular. Automated trading algorithms increased the value of the company to $5 Billion [3]. The price of shares in Cynk rose from six cents to $21 [4].
| |
| − | | |
| − | Although the actual number of Twitterbots on Twitter is difficult to estimate, a report by Twitter in 2014 estimated that 5-8.5% of Twitter accounts are Twitterbots [5]. The large number of Twitterbots on Twitter is partly due to the fact that they are easily constructed using the Twitter Application Programming Interface (API).
| |
| − | | |
| − | There have been efforts made to develop technologies to identify Twitterbots and remove them from Twitter. The aim of this project is to analyse these technologies. This project is relevant in a modern context due to the widespread adoption of social media. It is important for government organisations to consider the impact of Twitterbots used in information warfare and to consider strategies that can be used to mitigate Twitterbots.
| |
| − | | |
| − | The main outcome of this project will be the production of a thesis document that will provide an in-depth analysis of the current state of Twitterbot identification techniques. This document will be relevant to government organisations and other researchers to provide an overview of the current technologies and the possible direction of future research in Twitterbots.
| |
Revision as of 22:22, 5 December 2017
Automated accounts on Twitter known as Twitterbots are being used for malicious purposes. There is a particular concern in the use of Twitterbots in information warfare to influence politics or cause civil unrest. Research has been conducted into methods that could be used to identify Twitterbots with the aim of removing their presence from Twitter. The aim of this project is to evaluate the Twitterbot identification methods that have been proposed and to propose a new framework for Twitterbot identification. The techniques examined include human-based methods, supervised machine learning, activity correlation, monitoring user activity and clustering. In addition a botnet case study has been examined that shows that a targeted approach to Twitterbot identification can be effective
