Enhancing the Automated Analysis of Criminal Careers

Four enhancements have been devised, both on a semantic and efficiency level, that improve existing methods of automated criminal career analysis. A new distance measure was introduced that more closely resembles the reality of policing. Instead of previous, more rigid, comparison of career changes over time we propose an alignment of these careers. We employ a faster and better method to cluster them into a two-dimensional plane. This visualization can ultimately be used to predict the unfolding of a starting career with high confidence, by means of a new visual extrapolation technique. This paper discusses the applicability of these new methods in the field and shows some preliminary results.