ICAD: An Intelligent Framework for Real-Time Criminal Analytics and Detection.

Criminal investigation plays a vital role nowadays where the law enforcement agencies (LEAs) carry out this critical mission thoroughly and competently. However, such complicated mission involves a broad spectrum of tasks including collecting evidences from various data sources, analyzing them, and eventually identifying the criminals. Particularly, data may be collected by LEAs from telecommunication companies, online money transfer agencies, social media networks, video surveillance systems, bank transactions, and airways companies. LEAs confront various challenges from different fronts regarding criminal investigation. Thus, handling such big and heterogeneous data coming from different sources and recognizing potential suspects in a real-time is becoming a major challenge for LEAs in criminal investigation. In this paper, we propose an end-to-end Intelligent framework, called as ICAD, to help LEAs in Criminal Analytics and Detection. Mainly, ICAD uses cutting-edge technologies (data science and big data tools) as well as ontological models and inference rules to automatically identify suspects and reduce the human intervention in the investigation process. Furthermore, ICAD consists of four phases. The data sources phase in which we take benefits of various data collection sources that are essential in the crime investigation process. The data acquisition phase where data are collected, preprocessed, and stored using data science tools. The model phase in which a criminal-based ontology is defined that semantically integrates and enriches real-time data into useful information. The last phase is the knowledge extraction where a set of inference and reasoning rules are defined and applied over the ontology to detect criminals according to their activities.