Litcius/Paper detail

A dataset for the detection of fake profiles on social networking services

Samuel Delgado Munoz, Edward Guillén

202017 citationsDOI

Abstract

The use of multiple social media platforms is a common practice on more than two-third of all Internet users, according to OurWorld In Data. From this perspective, the verification of a real profile is a matter of growing interest, because false virtual identity could trigger problems such as spoofing, bots, grooming, sextortion, just to name a few. This paper presents a method to detect fake profiles on social media platforms by deploying some machine learning detection methods over a novel dataset. The dataset was designed with 17 metadata features from real and fake profiles and it was tested on Instagram profiles. After deploying different machine learning algorithms, the obtained detection rate was near to 96% with good false positive rates.

Topics & Concepts

MetadataComputer scienceSocial mediaSpoofing attackThe InternetPerspective (graphical)Identity (music)False positive rateIdentity theftMachine learningArtificial intelligenceWorld Wide WebComputer securityAcousticsPhysicsSpam and Phishing DetectionNetwork Security and Intrusion DetectionAdvanced Malware Detection Techniques
A dataset for the detection of fake profiles on social networking services | Litcius