Fuzzy association rule mining for community crime pattern discovery

Current manual inspection of crime data by analysts is limited, primarily due to the amount of data that can be processed concurrently and in a reasonable time frame. Further, complex relationships between various crime attributes can be overlooked by human analysts. Providing automated knowledge discovery tools becomes attractive to accelerate the efforts of local law enforcement. In this paper, we study the application of fuzzy association rule mining for community crime pattern discovery. Discovered rules are presented and discussed at regional and national levels. Rules found to hold in all states, be consistent across all regions, and subsets of regions are also discussed. A relative support metric was defined to extract rare, novel rules from thousands of discovered rules. Such an approach relieves the need of law enforcement personnel to sift through uninteresting, obvious rules in order to find interesting and meaningful crime patterns of importance to their community.