Home / Wiki / C / Credit Scoring Model

Credit Scoring Model

Compliance Risk and ControlsUpdated 16 Nov 2026

A credit scoring model is a statistical method that converts information about a person or a business into a number expressing how likely the borrower is to fail to repay. Lenders use the resulting credit score to accept or decline applications, set interest rates and credit limits, and monitor existing accounts. The best-known consumer example in the United States is the FICO score, built from records held by a credit bureau: payment history, amounts owed, length of history, new credit and the mix of accounts.

Traditional models are scorecards based on logistic regression, which are simple to explain and audit. Newer ones apply machine learning to wider data such as bank transactions and rent payments. Each estimates a probability of default. Models must be validated regularly, and under laws such as the Equal Credit Opportunity Act they may not discriminate on protected grounds, and lenders must be able to give reasons for a refusal.

Last updated 16 Nov 2026