Abstract:
Criminal intelligence (CI) is an analytic process of examining criminal data in search of patterns, relationships or networks between different crimes and its associates. The impact of impairing behaviours on health of the predisposed individual criminal results in patterns which need to be investigated intelligently for the discovery of pieces of knowledge required for preparedness, classification, predication and response to medical emergency cases. Since there is dearth in knowledge in this domain, and the failure of frequency and database methods in associating crimes with ailments and vice versa, handling medical conditions of criminals at an early stage becomes an arduous task. This paper advances CI by proposing an intelligent framework for mining criminal health data using link analysis, cluster analysis and support vector machine (SVM). The design incorporates a five-stage data pre-processing; a knowledge warehouse built from knowledge marts of criminal health in bottom-up manner. Fuzzy c-means splits the dataset and associated diseases to more than one crime and creates a fuzzy inference system utilized by SVM for classification and prediction of a criminal’s disease and vice versa. This work would assist stakeholders in the management and control of criminal activities and in the preparedness for criminal health emergencies as well as serving as a vital analytic and decision support model in criminal health behavior domain. The implementation of the design and assessment of its performance with data on criminal health activities is for further research.
Keywords: Association mining, criminal behavior, fuzzy c-means, criminal intelligence, link analysis, fuzzy c-means, pattern discovery
