Police Use Software To Predict Recidivism
Software from SPSS, employing data-mining technology, will be used by the National Offender Management Service to analyze coded and plain written text to predict the likelihood of reoffense. Data on a number of predisposing factors, including demographics, lifestyle, and behavior, can be gleaned through analysis of the various records kept on each offender.
The software searches for common themes to determine the chance of recidivism. “With over 400,000 held, we have access to vast amounts of information on each individual that we work with,” noted Philip Howard with the National Offender Management Service. “However, until we started working with SPSS, much of the uncoded data that was being collected couldn’t be taken into account when analyzing the danger to the community of offenders with particular characteristics.”
SPSS U.K. general manager Noel Coloe said the text mining solution came out of more than 25 years worth of research into natural language processing.
Abstracted by the National Law Enforcement and Corrections Technology Center(NLECTC) from VNUNet (09/08/05); Thomson, Iain .