Association rule mining based crime analysis using apriori algorithm

Abstract

Nowadays, criminal law enforcement is a crucial task due to the increasing of crime rates, limitation of manpower and lack of awareness from the local community. Historical data on crimes activities need to be analyzed in order to get the trend and pattern of crimes for future prevention actions. The aim of this article is to explore the relationship between the category of location and area of criminal records through extracting the patterns that frequently occur by applying Apriori algorithm from Association Rule Mining (ARM) method. CRISP-DM methodology is used to conduct this study that consists of business and data understanding, data preparation, modeling, evaluation and deployment phases. As a result, there are strong rules being created from the high support and confidence in modeling process where the generated rules would be considered as potential rules for pattern visualization in crime analysis. This study brings a high significance for effectiveness and efficiency strategy in criminal law enforcement and it can also be explored for other association rule mining methods for future work enhancement.