Bank Loan Repayment Prediction SystemCredit risk decision support

Factors considered by the model

The model analyses historical repayment patterns to estimate future loan repayment behaviour. Categorical attributes are one-hot encoded and aligned to the exact training column order before scoring.

Customer Income

Annual income in LPA, indicating capacity to service the loan.

Borrowed Amount

Requested loan exposure in LPA relative to income.

Credit Score

Bureau score between 300 and 900 reflecting credit history.

Loan Tenure

Repayment duration in months, from 12 to 240.

Existing Loans

Number of currently active loan obligations.

Employment Status

Salaried, Self-Employed or Unemployed.

Education

High School, Graduate or Post-Graduate.

Marital Status

Married or Single.

Age

Applicant age between 21 and 65 years.

How the prediction is produced
End-to-end flow used for every application.
  1. Step 1Collect customer input
  2. Step 2Validate every field against trained ranges
  3. Step 3One-hot encode categorical variables
  4. Step 4Align columns to the training schema (missing dummies set to 0)
  5. Step 5Score with the trained Logistic Regression model
  6. Step 6Return repayment / default decision with probabilities