A Framework to Assess Risk of Illicit Trades Using Bayesian Belief Networks

Recent years have seen the initiatives against illicit trades gain significant traction at both national and global levels. A crucial component in this fight is correct assessment of the risks posed by different trades across different regions. To aid in this cause, we provide a risk prediction framework based on Bayesian Belief Networks. It involves the development of a causal model incorporating variables related to the rise/decline of the illicit trade volume. The influence of these variables are determined by training on available data that are allowed to update over time. Implementation on a sample case study shows relatively low prediction accuracy of our model. Factors constraining its performance are analyzed and possible ways to avert them are discussed. We expect this framework to act as a decision support tool to the policymakers and strengthen them in the fight against illicit trades.