Create a Generalized Linear Model

Here are the steps to create a generalized linear model:
  1. From the toolbar, click Menu Arrow Down icon next to New Visualization Icon. Select Generalized Linear Model from the drop-down list.
  2. Minimize the Linear Regression visualization.
  3. From the Data pane, drag and drop the Age at Death variable into the Response field in the right pane.
  4. In the Data pane, hold down the Ctrl key, and select Blood Pressure Status, Cause of Death, Leaf ID (1), Sex, Smoking Status, Cholesterol, Height, Smoking, Weight, and Diastolic*Systolic. Drag and drop these variables onto the model pane.
    Generalized Linear Model Results 1
  5. Click the Properties tab in the right pane. The Distribution property enables you to specify the distribution of the response variable and to build a model based on that distribution. The default distribution is Normal.
    To determine whether the normal distribution applies to the response variable, click Show Actions Button in the Data pane, and select Measure Details. In the Measure Details window, select Age at Death.
    Age at Death properties
  6. Notice that Age at Death is not normally distributed and is slightly skewed left. Click Close.
  7. Although the distribution is not exactly Poisson, use the Poisson distribution for this example. For the Distribution property, select Poisson. Next, select Identity for Link function.
    Note: You are encouraged to repeat this example with different distributions and link functions and compare their performance and to familiarize yourself with SAS Visual Statistics.
    Generalized Linear Model Results 2
  8. Save the project.
Copyright © SAS Institute Inc. All Rights Reserved.