NCJ Number
120636
Journal
Journal of Police Science and Administration Volume: 16 Issue: 3 Dated: (September 1988) Pages: 210-216
Date Published
1988
Length
7 pages
Annotation
This paper describes an ongoing project aimed at developing an expert system for suspect identification in armed robbery incidents.
Abstract
A prototype of the system, named AREST (Armed Robbery Eidetic Suspect Typing), has been completed. The main task of the AREST project is the construction of a knowledge base. This was done by eliciting and representing formalized procedures (questionnaires and other interviewing mechanisms) for solving armed robbery crimes. The feasibility of the AREST system was tested using a prototype knowledge base covering traits that a suspect might have and that could be observed by an eyewitness. These traits included sex, race, amputations, deformities, facial scars, tattoos, facial oddities, eye defects, build, hair type, age, and height. A conventional computer program was developed to study the operation of the prototype classification system, and the prototype knowledge base was implemented in a composite structure where program code and data were intermixed. This was done to enable the researchers to study and understand the operation of a classification procedure prior to actually constructing the expert system. Using rules to form logical procedures, along with hypotheses and inferences, the knowledge base for AREST can be expanded and refined to include new rules as developments dictate. System testing can be done locally with the aid of law enforcement officers. The testing, verification, and validation of AREST will cover operational characteristics, consistency, completeness, precision, and relative effectiveness in comparison to conventional methods. The AREST system can help detectives and police officers solve routine offender classification problems and enhance and expedite the process of solving a crime. 5 references, 1 table, 6 figures.