A simulator for a simple bank using the LendingClub public loan data set
Download the LendingClub data file appl-accepted.csv
from the kaggle site
https://www.kaggle.com/denychaen/lending-club-loans-rejects-data
Use the notebooks in ./notebooks/
to prepare the data, build the classifier, and simulate the bank.
Some notebooks begin by loading a pickle file generated by the previous step -- these can be found in the box folder.
- Prepare data:
prepare-data.ipynb
- Experiment with training the classifier:
compare-training.ipynb
- Train the core features classifier used in the simulation:
train-on-core-features.ipynb
- Run the bank lending simulation:
bank-simulater.ipynb
"...premature optimization is the root of all evil..." Donald Knuth, 1974 (https://en.wikiquote.org/wiki/Donald_Knuth)