There has been a recent movement to apply business techniques to law enforcement and intelligence. According to an article in the July/August 2006 edition of IT Pro, analysts can actually apply specific statistical algorithms to historical data to predict crime in densely populated cities, such as NYC. Unfortunately, predicting low frequency events such as crime often has poor overall accuracy. It can however, be useful in determing the distribution of force deployment. Therefore advanced data mining and predictive analytics are now being used to create statistical models to determine where to deploy enforcement. The following figure illustrates an example of this type of application:
It's amazing to me how accurately we can predict outcomes using data mining. At what point do you think we are just writing algorithms to try to predict the unknown to simply make us
feel safer? Do we try to explain the future using a mathematical function so that we can simply
say we are cracking down on crime or can we
really use statistical models (such as the figure above) to improve public safety in areas such as NYC where location really matters?
No comments:
Post a Comment