Skip to content
/ loan Public

Predicting Default Borrowers using dataset from LendingClub.com.

Notifications You must be signed in to change notification settings

hybchow/loan

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 

Repository files navigation

Predicting Default Borrowers

Using the LendingClub.com dataset, we compared and critically evaluated the use of Multilayer Perceptron (MLP) and Support Vector Machine (SVM) for predicting default borrowers. The PyTorch and Scikit-learn packages in Python were used to develop the models.

Both MLP and SVM models exhibited good capability in detecting default borrrowers. Based on the AUC score, the SVM model was slightly more robust than MLP. The higher sensitivity of SVM (34%) compared to MLP (25%) also suggested the SVM model was more suitable for this task, as a high cost is associated with the incorrect classification of default borrowers.

About

Predicting Default Borrowers using dataset from LendingClub.com.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published