Slashdot Mirror
Software Used To Predict Who Might Kill
eldavojohn writes "Richard Berk, a University of Pennsylvania criminologist, has worked with authorities to develop a software tool that predicts who will commit homicide. I could not find any papers published on this topic by Berk, nor any site stating what specific Bayesian /
decision tree algorithm /
neural net is being implemented." From the article: "The tool works by plugging 30 to 40 variables into a computerized checklist, which in turn produces a score associated with future lethality. 'You can imagine the indicators that might incline someone toward violence: youth; having committed a serious crime at an early age; being a man rather than a woman, and so on. Each, by itself, probably isn't going to make a person pull the trigger. But put them all together and you've got a perfect storm of forces for violence,' Berk said. Asked which, if any, indicators stood out as reliable predicators of homicide, Berk pointed to one in particular: youthful exposure to violence." The software is to enter clinical trials next spring in the Philadelphia probation department. Its intent is to serve as a kind of triage: to let probation caseworkers concentrate most of their effort on the former offenders most likely to be most dangerous.
Here are the pertinent details:
Title: Forecasting Dangerous Inmate Misconduct: An Application of Ensemble Statistical Procedures
Journal: Journal of Quantitative Criminology
Issue: Volume 22, Number 2 / June, 2006
Pages: 131-145
Abstract:
In this paper, we attempt to forecast which prison inmates are likely to engage in very serious misconduct while incarcerated. Such misconduct would usually be a major felony if committed outside of prison: drug trafficking, assault, rape, attempted murder and other crimes. The binary response variable is problematic because it is highly unbalanced. Using data from nearly 10,000 inmates held in facilities operated by the California Department of Corrections, we show that several popular classification procedures do no better than the marginal distribution unless the data are weighted in a fashion that compensates for the lack of balance. Then, random forests performs reasonably well, and better than CART or logistic regression. Although less than 3% of the inmates studied over 24 months were reported for very serious misconduct, we are able to correctly forecast such behavior about half the time.
Unfortunately, you've got to pay $30 to get this paper. Maybe some slashdotter with a school/corp subscription to Springer will put up the text? ;-)
I thought Phillip K. Dick already explained why this was a bad idea...
There, I edited that for you buddy.
Let's just leave it at that's what you really intended, because otherwise I'll destroy all of my karma in spewing forth a slur of obscenities about how...
well, let's just leave it at that.
1) Convicted criminals are the only ones that concern probation officers.
2) Convicted criminals are the only ones they are likely to have the data to fill most of the fields for.
3) Probation officers have a job to do that does not involve tracking random citizens.
Thus, it seems unlikely it could be used for anything *but* the intended purpose without a fairly serious rework.
Integrate Keynote and LaTeX
It looks like Scotland Yard is also looking for scary new tactics in fighting crime. The latest idea of Laura Richards, head of analysis of the Metropolitan Police's Homicide Prevention Unit, sounds like a strangely familiar concept to those who have seen Minority Report. She aims to create a database of people who could supposedly commit a crime in the future, based on their psychological profile.
Even though preventing crimes is a noble motivation, this idea raises serious privacy issues.
As a sidemark it should be mentioned that Laura Richard also seems to be part of the team that "revealed" Jack the Ripper's face some time ago.