Application of a Profile Similarity Methodology for Identifying Terrorist Groups That Use or Pursue CBRN Weapons

摘要

No single profile fits all CBRN-active groups, and therefore it is important to identify multiple profiles. In the analysis of terrorist organizations, linear and generalized regression modeling provide a set of tools to apply to data that is in the form of cases (named groups) by variables (traits and behaviors of the groups). We turn the conventional regression modeling "inside out" to reveal a network of relations among the cases on the basis of their attribute and behavioral similarity. We show that a network of profile similarity among the cases is built in to standard regression modeling, and that the exploitation of this aspect leads to new insights helpful in the identification of multiple profiles for actors. Our application builds on a study of 108 Islamic jihadist organizations that predicts use or pursuit of CBRN weapons.