This module focuses on building regression models and selecting the best set of predictors using practical, data-driven methods in SAS. You’ll start by setting up the course environment, then move into key model selection approaches—including all-possible regressions, stepwise selection using significance levels, and selection using information criteria. Along the way, you’ll learn how to interpret p-values and parameter estimates, evaluate models with metrics like adjusted R-square and Mallows’ Cp, and apply these through demos and practice assignments.

In this module you learn to verify the assumptions of the model and diagnose problems that you encounter in linear regression. You learn to examine residuals, identify outliers that are numerically distant from the bulk of the data, and identify influential observations that unduly affect the regression model. Finally, you learn to diagnose collinearity to avoid inflated standard errors and parameter instability in the model.

In this module you learn how to transition from inferential statistics to predictive modeling. Instead of using p-values, you learn about assessing models using honest assessment. After you choose the best performing model, you learn about ways to deploy the model to predict new data.

In this module you look for associations between predictors and a binary response using hypothesis tests. Then you build a logistic regression model and learn about how to characterize the relationship between the response and predictors. Finally, you learn how to use logistic regression to build a model, or classifier, to predict unknown cases.