Bank Account Classification for Gambling Transactions

摘要

Financial services provided by banks are often misused for money laundering. Gambling is a criminal offense in the jurisdiction of the Republic of Indonesia. The increase in the number of customers is a challenge for the Bank to supervise the financial transactions. Based on the previous research, classification in machine learning often used to identify money laundering in the banking industry. However, there are no studies about identifying misuse of accounts in gambling activities using machine learning. This research used real financial transaction data using the SLR, experimental, and semi structured interviews with several Subject Matter Expertise. The results of this study indicate the classification with ensemble algorithms such as LightGBM can identify gambling accounts. Based on the evaluation results of classification performance with LightGBM, this model has the best performance compared to other ensemble models at a precision of 97.26%. These findings can be used as a basis for automatic reporting of suspicious financial transactions to financial regulatory.