Agent-Based Modelling: Geo-surveillance of individuals in a CCTV controlled area based on social media psycholinguistic profiling.

Abstract

The work presented in this dissertation revolves around data providers available for the collection of social media data. This data is then profiled, using psycholinguistic profiling, to determine ‘dark-traits’, such as psychopathic tendencies. The parsed data is then used in experiments conducted with a bespoke system created specifically for this research. The system models complex human behaviour, allowing agents to make decisions based on their surroundings and knowledge of other agents, including personality and existing connections (family or friends). The experiments recorded each agent’s geolocation to an external file which was then used to overlay the agent’s geolocations onto a map of the closed-circuit television (CCTV) system geolocations which was analysed to determine and observe how many ‘high-risk’ or ‘dark-trait’ agents were caught in the CCTV view-shed(s). The findings and new research into big data, psychology, social media profiling, and CCTV crime analysis within this project has led to my involvement in two publications.