A study of credit scoring using 3 types of classifiers: logistic regression, neural network & random forest, implemented as a ensemble learning hard voting classifier.
Everyone working on this project is a member of the NUS Fintech Society's Machine Learning Department.
Link to the article:
Using a dataset about loan transactions from 2009 to 2017 on the P2P platform Prosper, we set out to compare the 3 types of classifiers & their performance at performing a classification of whether a potential lender, based on the loan request issued, is either likely or unlikely to repay the loan.
| Category | File |
|---|---|
| EDA | feature_EDA_2.ipynb |
| Neural Network | NN_notebook final.ipynb |
| Random Forest | Forest.ipynb |
| Logistic Regression | LogReg.ipynb |
| Name | Role | GitHub Profile | Contributions to Source Code |
|---|---|---|---|
| Koby Chua | Tech Lead | @PopKobs | EDA, Random Forest implementation |
| Lee Wen Yeong | Neural Network Lead | @harvestingmoon | EDA, Neural Network implementation |
| Jerry Yang | EDA Lead | @mcxraider | EDA, Neural Network implementation |
| Tan Jing Jie | Neural Network & EDA co-lead | @jjtan444 | Neural Network implementation |
| Sparsh Kumar | Random Forest co-lead | @justsparsh | Random Forest implementation |
| Yang Yee | Random Forest co-lead | @yangyee-hub | Random Forest implementation & performance measures |
| Peh Ting Xuan | Logistic Regression co-lead | @tingxuanp | Logistic Regression implementation |
| Lionel See | Logistic Regression co-lead | @lionsee77 | Logistic Regression implementation |
Prosper Loan Data, Henry Okam, October 2022, https://www.kaggle.com/datasets/henryokam/prosper-loan-data