A group of computer scientists was able to develop a tool that can easily detect people who create fake profiles online, such as providing the wrong gender or age, which can bring about positive benefits to ensuring the safety of social networking sites.
The University of Edinburgh scientists trained computer models to recognize social media users who create false personal information and are regarded as catfishes.
They came up with computer models that identify fake profiles on an adult content website because they believe this type of site is a serious target of catfishes where they can befriend other users and increase profile views.
“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,” Dr. Walid Magdy, from the School of Informatics in said university, said in a statement.
The computer models were built on data collected from around 5,000 confirmed public profiles on the site.
The profiles were utilized for training said computer models in accurately estimating the age and gender of users through their writing style in comments section and network activity.
Hence, the models were able to accurately spot the misinformation and estimate the gender and age of users who use unverified accounts.
For protection of user privacy, all details remained anonymous.
Nearly 40 percent of users of the site invent their age, while one quarter falsify their gender, noting that women are more likely to dupe than men.
The findings highlighted the extent of catfishing activities on adult networks and showed the effectiveness of such technology in spotting dishonest users.
The study will be presented during the International Conferences on Advances in Social Networks Analysis and Mining in Australia.
The King’s College London, Lancaster University, and Queen Mary University of London contributed to the research.