Credit Default Modeling Using Machine Learning Algorithms
This project is submitted as a final project for Computational Statistics module. Because of the limitations in size, the .pdf file is meant to be short enough to show all the necessary parts of the project.
This project aims to compare the Machine Learning algorithms for measuring credit card default risk. I will directly use the data UCICreditCard. Further processing is carried out, and I will include those in the code scripts. The main idea of the project is to implement several Machine Learning algorithms in credit default context, and compare their predictive power, error rates and reliability in general. Some important measurements around Confusion Matrix will be discussed. Also, variable importance is considered across models to find the best fitting model and to decrease computational costs of the model implementation. Project also includes a simulation part where hypothetical data is generated and used for building models.