SIMBox fraud is one of the most prevalent scams in cellular networks that cause a significant monetary loss to mobile operators. It consists of diverting international voice traffic from regulated routes and rerouting it as local calls in the destination country. SIMBox fraud mitigation is challenging as it requires knowledge from both regular and fraudulent parties. This paper identifies and addresses such challenges through the realistic design of a simulator, i.e., FraudSIM, enabling the generation of datasets enriched with regular and multiple fraudulent strategies-induced behaviors.