Sunday, July 4, 2010

Predictve analysis emerges as promising crime fighter

Youngstown Vindicator
Sunday, July 4, 2010

As many states face a third consecutive year of budget deficits, an increasing number of state employees face layoff or even job loss, which not so long ago would have seemed as unlikely as a 138 game set at Wimbledon.

Included in those quickly vanishing state jobs are police officers. Pennsylvania Gov. Ed Rendell is set to leave office with fewer troopers on the payroll than when he took office. The Ohio State Highway Patrol Mission Review Task Force is looking at options, including staff reductions, to reduce the $319 million highway patrol budget. The problem is dire in other states. Illinois will lay off 460 state troopers. Michigan has eliminated 100 state police jobs. Tennessee has fewer troopers than in 1977.

Law enforcement agencies are looking for means, other than boots on the ground, to keep a handle on a crime rate that is at its lowest point in nearly 45 years. An area that is generating some attention is the development of cutting-edge crime prediction analysis.

Most police departments tout their clearance rates—the percentage of crimes solved by the department. The ability to predict crime brings law enforcement to a whole new level—stopping the potential criminal before he commits a crime. To carry out that mission, the police must have an idea when, where, and by whom a crime may be committed.

The concept of predictive analysis derives from sophisticated computer models that sort historical data, identify trends, make correlations, and fuse all the information together. Some agencies are cooperating with the private sector and academia to establish systems to support effective predictive analysis.


The Florida Department of Juvenile Justice is using predictive analysis to improve its screening and placement process. Using an analytics system developed by IBM, FDJJ will analyze key predictors such as past offense history, home life environment, gang affiliation and peer associations to better predict which young offenders are at a higher risk to re-offend.

With that information, FDJJ can more effectively place specific segments of juveniles into the best programs for rehabilitation. For example, juveniles identified as having a high risk to re-offend can be placed in a more restrictive environment where treatment can be intense. Young offenders with a low risk of re-offending can be assigned to less restrictive supervision.

Researchers at UCLA believe they have developed a math model to help police predict and eliminate emerging crime hot spots—areas that have an increasing likelihood for criminal activity. “We can actually define where you get hot spots and where you won’t,” Jeffrey Brantingham, a UCLA associate professor of anthropology who has been working to define crime patterns,said. The Los Angeles and Long Beach police departments have used UCLA’s work in predictive analysis.

The Ministry of Justice in the United Kingdom uses predictive analytics developed by IBM to assess the likelihood of prisoners re-offending upon their release to help improve public safety. With predictive technology, the Ministry of Justice is analyzing hidden trends and patterns within the data. Predictive analysis has helped identify whether offenders with specific problems such as drug and alcohol abuse are more likely to reoffend than other prisoners.

The Philadelphia Adult Probation and Parole Department collaborated with the University of Pennsylvania to establish a predictive tool for murder. Professor Richard Berk has focused on ways to distinguish probationers most likely to be charged with murder. According to Penn Professor Lawrence Sherman, Berk’s analysis can identify APPD probationers who are up to 42 times more likely, on average, to be charged with murder than other probationers. The model created by Penn forecasts homicide risk among individual probationers and parolees using statistical methods similar to those employed in hurricane forecasting.

Predictive analysis is no longer science fiction. The ability to predict crime hot spots and identify offenders with a high risk of re-offending has the potential to reduce crime, make neighborhoods safer and citizens less prone to victimization.

Find article here:

No comments:

Post a Comment