Commercial Bank Customers Gambling Money Laundering Risk Monitoring Based On Data Mining

With the development of the Internet in recent years,Internet gambling money laundering as a new form of money laundering has gradually attracted the attention of commercial banks’ AML monitoring and analysis staff.The explosive growth of customers suspected of Internet gambling money laundering poses difficulties for commercial banks’ AML monitoring and analysis.Commercial banks are actively seeking effective monitoring and analysis methods for Internet gambling money laundering in order to avoid omission and late reporting of suspicious transaction reports,as well as from the perspective of their own reputation risk.Based on data mining,the use of suitable machine learning algorithm models can effectively solve the problem of human and material resources strain faced by commercial banks in the process of Internet gambling money laundering risk monitoring..This paper analyzes the basic status of data mining in China and abroad,combines the anti-money laundering work situation at home and abroad,analyzes the data mining methods commonly used in daily work and their applicability,dissects the characteristics of Internet gambling money laundering,and finds that logistic regression,discriminant analysis,random forest and XGBoost models are applicable to the current anti-money laundering data monitoring of commercial banks.The data is then empirically analysed using logistic regression,discriminant analysis,random forest and XGBoost models after pre-processing the data with null and missing values,multiple covariance processing and variable significance analysis based on three months of customer data from a domestic commercial bank’s Chongqing branch.The model efficacy was judged by confusion matrix,model accuracy,recall and other indicators,and the model fit was visualised using ROC curves or 10-fold cross-validation results data.Through the comparative analysis of the empirical results,this paper considers that in the course of this commercial bank’s AML monitoring and analysis,the random forest model can effectively achieve the classification and prediction of internet gambling money laundering customers,which is more effective than the classification and prediction effects of logistic regression,discriminant analysis and XGBoost,and can provide reference standards and decision-making basis for commercial banks to better achieve the monitoring and analysis of internet gambling money laundering risks.