Dear Math Geeks,
Duncan spent last weekend (including Monday) doing some sort of Math Modeling Competition at which they used dating service algorithms to predict the likely next locations of serial crime.
According to what he told me last night, they can put in all the specifics of crime instances in a chosen area, and then use the characteristics of the community itself to model the likelihood of the next crimes in the series. I can't remember exactly what he said, as I was entering his FAFSA data at the time, but apparently using the dating service algorithm they can predict the next crime with a high degree of accuracy.
I presume that, unlike the random characteristic of humans looking for a mate, criminals are less likely to be as picky over a target location than two people deciding on whether to have a second date...
This sounds really reasonable to me -- how about those of you that understand how modeling works?
He's going to present on the results of the model at some conference in April, which since he is a junior is amazing to me! Such a proud Mama!
(The picture is my brain on high level math.)
According to what he told me last night, they can put in all the specifics of crime instances in a chosen area, and then use the characteristics of the community itself to model the likelihood of the next crimes in the series. I can't remember exactly what he said, as I was entering his FAFSA data at the time, but apparently using the dating service algorithm they can predict the next crime with a high degree of accuracy.
I presume that, unlike the random characteristic of humans looking for a mate, criminals are less likely to be as picky over a target location than two people deciding on whether to have a second date...
This sounds really reasonable to me -- how about those of you that understand how modeling works?
He's going to present on the results of the model at some conference in April, which since he is a junior is amazing to me! Such a proud Mama!
(The picture is my brain on high level math.)