README.md

---
license: mit
---

# Flawed Simulated Loan Approval Chance Predictor

This model exists to be used within a course to demonstrate model inversion attacks.

To interact with this model:
1. Download the .pkl file
2. In the same directory the .pkl file is in, create a python script.
3. Within the python file, include the following:
   ```python
   import pandas as pd
   import joblib

   #load the model with this function!
   def load_model():
     return joblib.load('model.pkl')

   model = load_model()

   #City_codes range from 0-4, Income and CreditScore fields are also required. 
   user_input = pd.DataFrame({'City_Code': 0, 'Income': 20000, 'CreditScore': 100}, index=[0])

   # predict the probability of a loan [[rejection, approval]] based on inputs 
   probas = model.predict_proba(user_input)

   # print the output. Will be in the form [[rejection chance, approval chance]]
   print(probas)
   ```
   4. Run the script to view the output. Will be in the form \[\[rejection chance, approval chance\]\]

   Leave questions in the community section.