Behavioural & Tempo-Spatial Knowledge Graph for Crime Matching through Graph Theory

Abstract

Crime matching process usually involves the time tedious and information intensive task of eliciting plausible associations among actors of crimes to identify potential suspects. Aiming towards the assistance of this procedure, we in this paper have exhibited the utilization of associative search; a relatively new search mining instrument to evoke conceivable associations from the information. We have demonstrated the use of threedimensional, i.e. spatial, temporal, and modus operandi based similarity matching of crime pattern to establish hierarchical associations among the crime entities. Later we used these to extract plausible suspect list for an unsolved crime to facilitate the crime matching process. A knowledge graph consisting of tree structure coupled with the iconic graphic is used to visualize the plausible list. Additionally, a similarity score is calculated to rank the suspect in the plausible list. The proposed visualization aims to assist in hypothesis formulation reducing computational influence in the decision making of criminal matching process.