Title：Credit Risk Modeling with Machine Learning 採用機器學習建立信用風險模型
Credit scoring is a useful policy-making method for banks to make decisions as to whether loans should be granted. The German credit
data set is a classical data set used to establish credit-scoring models. However, the classification accuracies have only been on
average, around 75%. In this talk, I will introduce state-of-the-art learners that have used to establish credit scoring models. If time
allows, I will discuss the possible reasons of why no breakthroughs have been achieved thus far in improving the classification accuracy.