Open social data crime analytics
Madden, Michael G.
MetadataShow full item record
This item's downloads: 349 (view details)
Ullah, Ihsan, Lane, Caoilfhionn, Drury, Brett, Mellotte, Marc, & Madden, Michael G. (2017). Open social data crime analytics Paper presented at the IWAISe: First International Workshop on Artificial Intelligence in Security: An IJCAI Workshop, Melbourne, Australia.
Crime is under-reported. Reporting crime requires the victim to complete a number of administrative obligations. These obligations, as well as the nature of the crime, may create an inertia that discourages the reporting of the crime (for example, being defrauded might damage a financial organisation s reputation). However, there may be information leaks from compromised organizations, via affected customers on social media. A key advantage of using social data is that it is often immediate, and can have indications of the nature of a crime such as (1) named entities, for example, Bitcoin or PayPal; (2) geocoding information; and (3) the affected persons. Our aim in this work is to use social media platforms e.g. Twitter, Reddit, Facebook, etc. to detect signals of cybercrime incidents. Such signaling is arguably a better indicator of the extent and effect of cybercrime than traditional reporting methods.