Around 50% of the world population is now active on internet, often subscribing websites, social networks or other online services. In this scenario, the issue of online faked identities is more and more present, with the phenomena of identity alteration, identity theft and identity fraud. To date, there are no systems able to detect people who subscribe or authenticate an online service with faked personal information. Moreover, the existing validated lie detection techniques are not suitable to be applied in the online environment. Starting from a previous study, this paper investigates the possibility to detect faked identities recording keystroke dynamics, while the user is filling an online subscription form with personal – real or faked, information. Cognitively overloading liars through few unexpected questions, we demonstrated that it is possible to recognize the deceivers with an accuracy of 85%. To automatically detect liars, three machine-learning classifiers were trained on 40 liars and 40 truth-tellers, and tested on 10 unseen liars and 10 truth-tellers. Liars have proved to be distinguishable from truth-tellers as they make more errors and are slower in typing unexpected information about their identity.

The online identity detection via keyboard dynamics

Monaro M.;Spolaor R.;Li Q. Q.;Conti M.;Gamberini L.;Sartori G.
2019

Abstract

Around 50% of the world population is now active on internet, often subscribing websites, social networks or other online services. In this scenario, the issue of online faked identities is more and more present, with the phenomena of identity alteration, identity theft and identity fraud. To date, there are no systems able to detect people who subscribe or authenticate an online service with faked personal information. Moreover, the existing validated lie detection techniques are not suitable to be applied in the online environment. Starting from a previous study, this paper investigates the possibility to detect faked identities recording keystroke dynamics, while the user is filling an online subscription form with personal – real or faked, information. Cognitively overloading liars through few unexpected questions, we demonstrated that it is possible to recognize the deceivers with an accuracy of 85%. To automatically detect liars, three machine-learning classifiers were trained on 40 liars and 40 truth-tellers, and tested on 10 unseen liars and 10 truth-tellers. Liars have proved to be distinguishable from truth-tellers as they make more errors and are slower in typing unexpected information about their identity.
2019
Advances in Intelligent Systems and Computing
Future Technologies Conference, FTC 2018
978-3-030-02682-0
978-3-030-02683-7
File in questo prodotto:
File Dimensione Formato  
monaro2018.pdf

non disponibili

Tipologia: Published (publisher's version)
Licenza: Accesso privato - non pubblico
Dimensione 523.95 kB
Formato Adobe PDF
523.95 kB Adobe PDF Visualizza/Apri   Richiedi una copia
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11577/3340665
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 7
  • ???jsp.display-item.citation.isi??? 5
  • OpenAlex ND
social impact