Data Mining Using Sas Enterprise Miner: A Case Study Approach by Sas Inst

Data Mining Using Sas Enterprise Miner: A Case Study Approach by Sas Inst

Author:Sas Inst
Language: eng
Format: epub
Tags: Team DDU
Published: 0101-01-01T00:00:00+00:00


60

Evaluating the Model

Chapter 2

Consider the choices under Number of Variables in the lower-right corner of the window. The items here enable you to select a specified number of effects to begin the selection process (for forward) or select a minimum number of effects to remain in the model. The order depends on the order that is displayed in the Interaction Builder. To change the order of effects, you can select

Tools

Model Ordering

but no ordering is done for this example.

Stepwise Stopping Criteria enables you to set the maximum number of steps before the Stepwise method stops. The default is set to twice the number of effects in the model. Stop enables you to set the maximum (for Forward method) or minimum (for Backward method) number of effects to be included in the model.

The Stepwise method uses cutoffs for variables to enter the model and for variables to leave the model. Changing these values might affect the final variables that are included in the model. In this example, clear the selection of the Defaults check box, change the value of Stepwise Stopping Criteria to 14, and change the value of Stay Significant Level to 0.025.

Close the Regression node and save the changes when you are prompted. Since you

have changed the default settings for the node, you will be prompted to change the default model name. Type StepReg for the model name.

Select OK .

Evaluating the Model

Right-click the Assessment node and select Run. This enables you to generate and compare lift charts for the two regression models. Observe that each node becomes green as it runs. Since you ran the flow from the Assessment node, you are prompted to see the Assessment results. Select Yes to see these results. Inspect the table that appears. If you scroll the window to the right, Enterprise Miner displays several typical regression statistics such as asymptotic standard error (ASE).



Download



Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.