Loan Approval Predictor (BUS458 Final) — kmsmit49

How it works

• The app tries to load loan_model.pkl from this Space repository.
• If the model is not present, you can upload a compatible loan_model.pkl file using the in-app file uploader.
• The model should expect the following numeric features, in any order but with these column names:
  • Granted_Loan_Amount
  • Requested_Loan_Amount
  • FICO_score
  • Monthly_Gross_Income
  • Monthly_Housing_Payment

Files

• app.py — Gradio application
• loan_model.pkl — (You must upload this file to run predictions)
• requirements.txt — Python dependencies

Notes

• Save the full preprocessing + estimator as a scikit-learn Pipeline to avoid mismatches.
• If loan_model.pkl is large (>50 MB), use Git LFS or host the model in a separate Hugging Face model repo and change the loading code accordingly.