Taiwan Credit Default Risk Modeling

OBJECTIVE

Develop a repeatable analytical workflow for predicting whether a customer may default, while considering the practical tradeoff between identifying risk and limiting false alarms.

APPROACH

Prepared the Taiwan credit-card default dataset, explored candidate predictors, built multiple classification models in SAS Enterprise Miner, and compared individual and ensemble approaches using model-evaluation measures.

ANALYTICAL VALUE

Demonstrated how predictive models can support credit-risk prioritization, portfolio monitoring, and more consistent decision-making when model results are interpreted alongside business costs.

SKILLS SHOWN

Classification modeling, model comparison, ensemble methods, risk interpretation, data preparation, and communicating predictive results.

TOOLS