Research On The Risk Of Cross-border Money Laundering And Terrorism Financing In Bank Card Industry

In terms of the increasing global money laundering(ML)and terrorist financing(TF)risks,FATF has published new evaluation methodologies with 40 recommendations since 2015,and members of FATF have strengthened their anti-money laundering(AML)and counter-terrorist financing(CTF)regulatory requirements to implement the risk base concept.Bank card payment is one of the most important tools for cross board settlement.Its AML/CTF related requirements are facing multiple challenges,i.e.volatile external environment,rapid technological innovation,etc.As the settlement switch institution,card association acts as an interchange for all transactions in the overall payment network,and is able to play a more important role in bank card cross-border risk prevention and control by taking its advantages in data capacity.Given the massive transaction data processed by card association,it will be more difficult to fully exploit and utilize the data's value via traditional control system,which is based on database and monitoring rules.This paper expounds the principle of Decision Tree,Random Forest and classifier performance evaluation index.After variables interpretation,data cleaning and exploratory data analysis,the C5.0 algorithm,CART algorithm and Random Forest are used to establish model on training data.Then it compares the performance of 3 classifiers and the generalization ability on test data.The example shows that,the integrated learning method(C5.0/random forest)is superior to the basic method(CART)in the detection of AML/CTF risk of cross-border transactions,and the 3 kinds of classification are applicable to anti-money laundering model.Through the empirical research into the AML/CTF risk of current cross-border transactions,the paper aims to enhance the efficiency and accuracy of the risk detection of the bank card cross-border capital.