Title

Real Time Detection of Social Bots on Twitter Using Machine Learning and Apache Kafka

Document Type

Conference Proceeding

Source of Publication

2021 5th Cyber Security in Networking Conference (CSNet)

Publication Date

10-14-2021

Abstract

Social media networks, like Facebook and Twitter, are increasingly becoming important part of most people's lives. Twitter provides a useful platform for sharing contents, ideas, opinions, and promoting products and election campaigns. Due to the increased popularity, it became vulnerable to malicious attacks caused by social bots. Social bots are automated accounts created for different purposes. They are involved in spreading rumors and false information, cyberbullying, spamming, and manipulating the ecosystem of social network. Most of the social bots detection methods rely on the utilization of offline data for both training and testing. In this paper, we use Apache Kafka, a big data analytics tool to stream data from Twitter API in real time. We use profile information (metadata) as features. A machine learning technique is applied to predict the type of the incoming data (human or bot). In addition, the paper presents technical details of how to configure these different tools.

Publisher

Institute of Electrical and Electronics Engineers (IEEE)

Volume

00

Disciplines

Computer Sciences

Keywords

Training, Social networking (online), Voting, Unsolicited e-mail, Blogs, Tools, Feature extraction

Indexed in Scopus

no

Open Access

no

Share

COinS