Skip to content

With historical data on loans and information on whether a borrower defaulted (charge-off) or not, can we build a model that can predict whether a borrower will repay their loan? Thus, in the future when a new prospective customer comes along, we can assess whether they are likely to repay their loan.

Notifications You must be signed in to change notification settings

AinulMr/Lending_Club_Keras_Project

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 

Repository files navigation

Lending Club - TensorFlow Keras Project


Ainul Marzuki Febri

Data: Lending Club

Lending Club is a peer-to-peer lending platform founded in San Francisco in 2006 by Renaud Laplanche. The platform connects borrowers seeking personal or business loans with investors looking to invest in those loans. Borrowers apply for loans online, and investors can choose to fund a portion of various loans to spread the risk. Lending Club enables easier access to credit for borrowers and attractive investment opportunities for investors.

Our Goal

With historical data on loans and information on whether a borrower defaulted (charge-off) or not, can we build a model that can predict whether a borrower will repay their loan? Thus, in the future when a new prospective customer comes along, we can assess whether they are likely to repay their loan.

Conclusion

This project successfully built a binary classification model to predict whether loans in Lending Club will be fully paid (class 1) or charged off (class 0) with 89% accuracy. Through the stages of data exploration, preprocessing (addressing missing data and handling categorical variables), and overfitting using techniques such as dropout and early stopping, the model showed excellent performance in detecting fully paid loans (class 1) with perfect recall (1.00) and high precision (0.88). However, the model had difficulty in detecting charged off loans (class 0) with a low recall (0.44), despite a very high precision (0.99). This suggests that while the model is effective in predicting fully paid loans, further improvements are needed to enhance the detection of charged off loans.

Recommendations for further improvement include:

  • Balancing the dataset if there is an imbalance between classes.
  • Exploration of additional features that can help detect class 0 better.
  • Using oversampling or undersampling techniques to overcome class imbalance.

About

With historical data on loans and information on whether a borrower defaulted (charge-off) or not, can we build a model that can predict whether a borrower will repay their loan? Thus, in the future when a new prospective customer comes along, we can assess whether they are likely to repay their loan.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published