Mining crime information
Using data mining techniques to help fighting crime sounds good. It is certainly an interesting topic. Can you imagine saying that you have caught a criminal using a decision tree? :-) Although this view is very simplistic, data mining seems to be helpful in some situations as pointed out by LocalTechWire. The main idea is to use data mining methods to identify crime trends and then anticipate crimes. This is certainly a trendy topic for data mining, proved by the publication of a new book by Colleen McCue: Data Mining and Predictive Analysis: Intelligence Gathering and Crime Analysis.
5 comments:
At KDML-2006 (Knowledge Discovery, Data Mining, and Machine Learning), Shyam Varan Nath from Oracle gave a talk titled Crime Patterns Detection Using Data Mining Framework (see second talk in session 2), in which he described an application of data mining to supporting police officers in detecting crime patterns and in thereby better solving crimes.
The paper for this talk will be made available online as part of the LWA-2006 proceedings. Please contact the LWA-2006 organizers for the details on when and where these proceedings will be available online.
One general aspect of law enforcement administration is the management (as opposed to elimination) of crime. Police forces already rotate their efforts to keep criminal activity from festering in one area for too long. Conseuently, data mining efforts to predict crime patterns do not need to be even close to perfect.
The police force of a major American city developed a neural network to assess police officers, based on complaints made against them by the public. I was given to understand that the system was fairly accurate in predicting the need for intervention, but the vendor in question told me that the police union shot it down.
-Will Dwinnell
http://will.dwinnell.com
The city of Richmond, VA has also implemented a data mining solution more for the purpose of manpower management (for example, when there is a high school football game between rivals, ...). They have presented at the past two SPSS Directions conferences (they are getting a lot of visibility in SPSS circles), but it isn't clear to me if it is really being used or is still in "cool technology" mode. http://www.spss.com/success/pdf/CS%20-%20Richmond%20PD%20LR.pdf
I have also spoken with a Southern California police department about identifying at-risk officers from the data they now collect about them. There are very clear hurdles with the unions on this issue, so I think Will's point is well taken. It's obviously not just about the data!
Will Dwinnell is correct re the management of crime. The work in Richmond is based on a "risk-based" deployment strategy, which is very similar to "just-in-time" policing.
The deployment study mentioned by Dean Abbott represents only a small fraction of the work that was done in Richmond, which also included motive determination modeling and the analysis of violent crime, as well as some homeland security-related analyses. This work has been replicated and extended to other areas of the country since that study.
Kelly McCue
I have also met Dr. Oatley from Sunderland univ. in a conference. He is applying data mining to criminal data. One of his paper can be found here.
Post a Comment