Developing Anti Money Laundering Identification using Machine Learning Techniques

Abstract

There are several areas that can be applied the machine learning techniques, one of which is the finance and economics sector, it can effectively help and assist on decision support on fraud investigation. In this sector the skills and methods of fraud is becoming more complicated. In financial sector there are a lot of fraud mechanisms from them money laundering is crucial activity to fraud. The main purposes of this study develop and recommend a machine learning algorithm for identification of Anti-Money Laundering. The bank data set collected from Commercial Bank of Ethiopia which has 14740 with number of missing value and then after effectively pre-processing the data set, the five different types of machine learning algorithms were used, these are support vector machine, Logistic Regression, Naïve Bayes Decision tree, and Random forest. From those machine learning algorithms, Random Forest the best performance with 99.1% accuracy.