Using Natural Language Processing to Map Police Interactions With the Severely Addicted and/or Mentally Ill Population

• In Event: Future Directions in Canadian Policing

Abstract

A sizable quantity of information within policing records management systems (RMS) remains as free-flowing natural language text. These narratives are often used by police services in small batches however, unlike pre-set categories within the RMS (e.g., time of event) these qualitative data are often not coded. As a result, the information contained within narratives cannot be reliably accessed computationally. Automated applications that can access this information may improve the efficiency and effectiveness of retrieving police service data. One area where this application may be useful is in the natural language text-found in RMS for police interactions with persons with severe mental illness (SAMI). This study uses a form of syntactic analysis known as natural language processing to computationally extract the 1) keywords and phrases and to 2) analyse the meaning (the semantics) from free-text police synopses that involved with SAMI. Results to be discussed.

Authors

• Adam Vaughan, Simon Fraser University

• Bryce Evan Stoliker, Simon Fraser University

• Martin A Andresen, Simon Fraser University

• Simon Verdun-Jones, Simon Fraser University

• P Brantingham, Simon Fraser University