Money Laundering Detection and Mitigation Using Decision Tree Ensembles

Money laundering is an unlawful activity that has caused underdevelopment to Africa, it's a ravaging vices that cripples economic growth, increases criminality and leads to economic sabotage. Many works previously done on money laundering detection, it was observed that it takes times before fraudulent transactions can be detected, it is also less effective, meaning that is can easily be circumvents and the data sampling methods used in previous researches makes the model to suffer from overfitting and loss of information, however this research adopts SMOTE (Synthetic Minority Oversampling Techniques) algorithm method to balance the dataset so as to have a fair representation of the two classes in order for machine learning model to accurately evaluate the models. The Objective of this research is to adapt and implement decision tree ensemble learning algorithms to detect money laundering activities, also design and implement an Anti Money Laundering (AML) query system to mitigate money laundering activities based on detection results and finally to carry out the performance evaluation of decision tree ensemble algorithms using standard performance metrics.