Can Concepts Reveal Criminals?

Abstract

In 2005 the Amsterdam-Amstelland police introduced Intelligence-led Policing as a management paradigm. The goal of ILP is to optimally use the information which becomes available after police patrols, motor vehicle inspections, video camera recordings, etc. to prevent crimes where possible and optimally allocate available resources. This policy has resulted in an increasing number of textual reports, video materials, etc. every year. Until now, this vast amount of information was not easily accessible because good analysis methods were missing and as a consequence hardly used by the criminal intelligence departments. In the first part of this talk I will give a short overview of traditional statistical methods such as hot spot analysis which have been used to make this information accessible and steer police actions. In the second part of this talk I will present using some real life cases how FCA was used to identify criminals involved in human trafficking, terrorism, robberies, etc. In the third part of this talk I would like to evoke a lively discussion on the potential of FCA related algorithms and methods for analyzing textual reports, internet data such as twitter feeds, browsing behavior of visitors of radical Islamic websites, etc.