Application of Data Mining In Forecasting NPLS Status of GSB Loan Case study in Chiang Mai University branch

  • ปารดา ศัสตุระ
Keywords: Machine learning, NPLs


Government Saving Bank (GSB) implemented the emergency Covid-19 relief loan policy to help individuals experiencing financial burden. The impact on the policy lead to NPLs. The purpose of this study are to study the parameters that lead to NPLs and to forecaste NPLS Status of GSB Loan using data mining. The technique in this study is the Cross-Industry Standard Process (CRISP). There are 6 processes that are business understanding, data understanding, data preparation, model development, evaluation, and deployment. The data collected from COVID-19 relief loan from 7,215 customer GSB Loan in Chiang Mai University branch during July-December 2020. To do the data preparation, feature selection for data mining i.e., data transfer, data reduction, data cleaning using WEKA as tool. There are 3 algorithms for forecasting NPLS that are Decision Trees, Naive Bayes and XXXXX. Algorithm performance testing by using the root of the mean square error (Root Mean Square Error : RMSE) for the best algorithm. The results can demonstrate that there are 5 parameters (occupation, status, education, income and gender) that lead to NPLs and  the most accuracy algorithm is Gradient Boosted Tree with accuracy 96.15 % comparing with Simple Regression and Random Forest with accuracy 94.52 % and 96.10 %, respectively.