Israeli Researchers Develop New Method for Detecting Fake Accounts on Social Media Platforms


Many if not most of the world’s most prominent social media platforms are riddled with countless numbers of fake accounts and bots, which impact not only upon the quality of use and reliability of information for users, but also the financials and trustworthiness of the platform and larger company. With ‘fake news’ being the buzzword of the moment and allegations of political interference running rife, there has never been a more pressing need to mitigate this ongoing issue.

Recognising the need for substantial action, a team of researchers from Israel’s Ben-Gurion University of the Negev and University of Washington, Seattle recently came together to develop a brand new method for detecting fake accounts on social networks such as Facebook and Twitter. Dima Kagan of the Department of Software and Information Systems Engineering at Ben-Gurion University, who served as lead researcher on the study, stated in no uncertain terms that “rooting out fake users has never been of greater importance.”

The new method, published in the latest issue of Social Network Analysis and Mining, works based on a simple assumption; that fake accounts tend to establish “improbable links” to other users in the networks. Building from this foundation the process then comprises two main ‘iterations’ (defined as a set of repeating processes, utterance or commands) which each derive their basis from machine-learning algorithms.

The first iteration is used to build a “link prediction classifier”, which subsequently provides an accurate estimation as to whether or not two particular users were legitimately linked. The second then uses this data to create new meta-features which researchers then incorporate into a generic classifier which serves to detect fake profiles.

As it stands there are reportedly more than 200 million fake accounts on Facebook alone, with Twitter harbouring an estimated 48 million on their own platform. In Twitter’s case this constitutes around 15% of their active user base. With this in mind users, investors and advertisers alike will be hoping this study bears some meaningful fruit and can make some form of impact. The researchers tested the algorithm on 10 different social networks and it apparently “performed well” on both simulated and real-world situations, outperforming other similar tools and showing the potential to enhance cyber security applications, so prospects sounds good at present.

Post a Comment

[blogger]

Author Name

Free Gift

Free Gift
Get immediate access to our in depth video training on the click by click steps required to get your successful online business started today

Contact Form

Name

Email *

Message *

Powered by Blogger.