Retrieval of criminal trajectories with an FCA-based approach

摘要

In this paper we briefly discuss the possibilities of Formal Concept Analysis for gaining insight in large amounts of unstructured police reports. Wepresent a generic human centred knowledge discovery approach and showcase promising results obtained during empirical validation. The first case study focusses on distilling indicators for identifying domestic violence from 4814 reports with the aim of better recognizing new incoming cases. In the second case study we used FCA in combination with Temporal Concept Analysis toidentify and investigate human trafficking suspects extracted from 266157 short observational reports. The third case study we present in this paper describesour application of FCA for identifying radicalising subjects from 166577 observational police reports. Finally, we conclude our paper with the case study on pedophile chat conversation analysis and the CORDIET data mining system.