What is Regression Analysis?

• A. Regression to the mean
• B. Identify which combination of attributes best predicts potential problems
• C. Of many attributes, only a few may be meaningful
• D. Most predictive attributes might be given more weight

Answer: (B)

Regression analysis is a statistical method used to quantify how an outcome variable changes when one or more predictor attributes change, so you can predict the outcome for new cases and understand which factors drive it. In the context of potential problems, it focuses on identifying which combination of attributes best predicts those problems.

By modeling the relationship between risk (the dependent outcome) and multiple borrower or financial statement attributes (the predictors), you get coefficients that show how each attribute relates to risk and how they work together. The model’s fit indicates how well those attributes together explain the risk, which helps you build a predictive scoring approach. This approach naturally weighs several factors at once, rather than relying on a single variable.

Other ideas described in the options—regression to the mean is about extreme observations moving toward the average, feature selection about only a subset of attributes being meaningful, and weighting attributes purely by perceived importance—are related concepts but don’t define regression analysis itself.