Email
Password
Remember meForgot password?
    Log in with Twitter

article imageUsing machine learning to detect fake on-line profiles

By Tim Sandle     Jun 20, 2017 in Technology
Edinburgh - The era of ‘fake news’ extends to ‘fake profiles’ on social media designed for the spreading of misinformation, as well as being used for criminal activities. A new computer program shows success in spotting these.
While social media outlets have taken strides to address fake news (Google and Facebook recently banned the use of their advertisement services on websites that post fake news), it is thought that many social media profiles are fake (both on the common sites like Facebook, as well as on other social networking sites). Such spoofed accounts can be anything from someone who’s using a made up name and image, all the way to scammers who create fake celebrity accounts.
The development comes from the University of Edinburgh and the aim of the computer software is to more readily identify people who use fake profiles online. This is through using machine learning to ‘train’ computer models to detect social media users who invent information about themselves. The term for these is “catfishes.”
One success, in trials, is with identifying users who are dishonest about their age or gender. Some predatory males, for example, may stalk young females by changing their gender or lying about their age. Foremost in the minds of the computer scientists who invented the software is safety on social networks.
The software has been tested out on an adult content website. The computer model was designed on information gleaned from about 5,000 verified public profiles on the adult site. The profiles were used to train the model to more accurately assess the gender and age of each user. This was based on collecting and analyzing information relating to style of writing and network activity.
An image representing online dating
An image representing online dating
UrbaneWomenMag
The study found that some 40 per cent of the adult site's users had given false information about their age. Also of interest was the fact that around 25 percent of the users had given the wrong gender. The technology would therefore be successful in weeding out dishonest users if it was applied to other sites.
Commenting on the study, lead researcher Dr Walid Magdy stated: "Adult websites are populated by users who claim to be other than who they are, so these are a perfect testing ground for techniques that identify catfishes. We hope that our development will lead to useful tools to flag dishonest users and keep social networks of all kinds safe."
The study was presented at a recent meeting of the International Conference on Advances in Social Networks Analysis and Mining in Australia.
More about machine learning, Artificial intelligence, Cybersecurity, Social med, fake news
More news from
Latest News
Top News