A Graph Mining Approach to Detect Sandwich Calls

Across the broad reception of numerous measures to detect fraudulent communication through phone calls in the context of crime investigation, the analysis of Call Detail Records among the suspicious group of individuals help to detect illegal activities based on their relationships. Criminologists and examiners make use of the network graph analysis methods to deal with these scenarios in a significant and productive manner. ‘Sandwich Call’ is a technique of making a chain of phone calls through a network of people to pass a message between two persons disguising the fact that they are effectively communicating with each other. Mostly it is used with an immoral intention as the resultant Call Graph of such communication does not reveal any sort of direct relationships between the entities easily. This work presents a Sandwich Call Detection Tool - a standalone graph mining-based tool to provide visual data representation and identification of Sandwich Calls. Its adoption may help in unveiling the structure of a criminal network and the roles and dynamics of communications among its components in real life.