Skip to content
Contact UsStart Building

Banking & Credit

Credit-card payment default

Estimate next-month payment default from billing and repayment history for credit-risk review.
AI-generated representative illustration of a single unbranded silver payment card beside a plain monthly statement envelope.
AI-generated task illustration. Representative scene; not a customer site or benchmark observation.
From known examples to a prediction

Available inputs

  • Credit limits and recent bill amounts
  • Repayments and payment-status history

Known examples

30,000 recorded examples

Prediction target

Default on the following month’s payment
0.61 pphigher ROC AUC

vs. LightGBM (tuned ensemble), the strongest conventional result in this benchmark.

Credit card payment defaultROC AUC · Higher is better

30,000 examples · 23 inputs · 9 evaluation splits

01
  1. Causilo
    0.7915
  2. LightGBM
    0.7855
  3. XGBoost
    0.7852
  4. CatBoost
    0.7847
  5. Extra trees
    0.7820
  6. Random forest
    0.7803
  7. K-nearest neighbors
    0.7691
  8. Logistic regression
    0.7454

The task in practice

The study pairs credit limits, bills, payments and recent payment-status history with next-month default labels. Each row represents a client’s recorded account history.

  1. Connect account histories with outcomes. Use recent account measurements and their next-month default labels.
  2. Describe another payment history. Supply the same limit, billing and repayment fields available before the outcome.
  3. Evaluate next-month risk. Check ranking and calibration against independently observed payment outcomes.

Scope. A recorded default signal does not establish the effect or appropriateness of a credit-limit change.

Dataset, sources and evaluation

Credit card payment default dataset. Original source and curated OpenML task linked below. Original dataset source.

The TabArena-curated OpenML version (46919) supplies the dataset and recorded target, DefaultOnPaymentNextMonth. Comparison scores come from the retained results_per_split_official.csv evaluation snapshot. These pages describe the task and compare the recorded means; source records are not republished here.

ROC AUC measures how well the scores rank the recorded positive outcomes above negative ones. A higher score is better. The change is reported in percentage points, rather than relative error reduction.

Causilo uses its default configuration. Each conventional family uses its strongest reported mean across default, tuned and tuned-ensemble settings. The chart compares reported means, without a claim of statistical significance, customer deployment or operational savings.

ROC AUC difference in percentage points = (Causilo mean − strongest conventional mean) × 100. ROC AUC is a ranking metric, not a percentage of correct predictions.

How to read these comparisons · All configurations and original values.