Regression and model assessment

The first method that we will use for fitting models to our data are regression models. This includes classical multiple linear regression models when the response is continuous. In our loan example the response is binary (BAD or NOT BAD) so we use logistic regression. Professor Joseph L. Schafer of Penn State University has a very nice PDF Introduction to Logistic Regression. The introduction includes some of the technical details about how to calculate the maximum likelihood estimates for a logistic regression model. Regression methods are applied using the Regression node.

We use model assessment procedures to see how the model performs on data sets different that the one used to fit the model. We do this with the Assessment node.