The Cyber Sleuth
DePaul University computer scientist Tom Muscarello has developed the Classification System for Serial Criminal Patterns (CSSCP), a computer system designed to help police investigate violent crimes, especially serial crimes.
Due to go live as early as next year, CSSCP employs pattern-recognition software to aggregate crime data, such as the age, gender, and height of the perpetrator, and the types of weapons and vehicles used, to generate a criminal profile. The system aims to overcome the trouble humans have finding patterns to connect serial crimes, a principal shortcoming in detective work.
CSSCP can sift through thousands of criminal reports each second, scanning for repeated phrases or words that would go unnoticed by human detectives. Serial crimes are particularly difficult to solve because they often happen over a long period of time and across a large area, sometimes involving police of different jurisdictions.
The system was patterned on the techniques of six of Chicago’s best detectives, said Muscarello, who has been working on artificial intelligence technology to help police since the mid 1990s. Built on the Kohenian neural network, the CSSCP system found at least 10 times the number of linked crimes as a group of detectives working with the same data in a recent trial.
While other computer-based detective aids exist, “nobody today in 2005 has come up with a program that has done what this network can accomplish,” said Charles Padgurski, formerly of the Chicago Police Department.
Abstracted by the National Law Enforcement and Corrections Technology Center(NLECTC) from the Chicago Tribune (11/29/05); Jones, Patrice M.