Loan Application Predictor

This app lets you input data into an AzureML neural network model trained on a Kaggle Dataset. Enter values manually, or use the "Test with Example Data" button to populate fields with random values. You can also click "Load PIF Example" (Paid in Full) or "Load Defaulted Example" to load preset data for each case.

*
*
*
*
*
*
*
*
*
*
*
*

Prediction Result

Probability of PIF: Not predicted yet

Example ID: None

The F1 score is a single number that balances how many good loan approvals you make (precision) with how many good applicants you correctly approve (recall), resulting in an optimal cutoff value of 63%.

Threshold Optimization Plot

Optimum cutoff value of 63%

SHAP Impact Plot

SHAP analysis plot.