Chapter 4 Questions
pdf
keyboard_arrow_up
School
University of Louisville *
*We aren’t endorsed by this school
Course
615
Subject
Information Systems
Date
Apr 3, 2024
Type
Pages
3
Uploaded by cecilal
4-44 Lesson 4 Models with Categorical Targets @ Practice 'S In this practice, you continue to use the PVA data set. You build a logistic regression model to classify those customers who donated. 1. Building a Logistic Regression in SAS Visual Statistics a. Return to your remote desktop client machine. If your session timed out, sign in. Use Student as the user ID and Metadata0 as the password. b. Start Visual Analytics or start a new report. Then select and open the PVA data source. ¢. Add a logistic regression to the canvas. d. If you did not do so already, in the Measure column, edit Target Gift Flag. Select Category to create a binary target variable for donations. e. From the menu bar, click : (Menu) = Interface options = Disable auto-refresh. f. Add Target Gift Flag as the response. g. Add Gender, Home Owner, and Status Category 96NK as classification effects. Then add all 23 variables as continuous effects except Target Gift Amount, Target Gift Amount with Zero, and Median Home Value Region. (You add 20 columns.) h. Inthe Options pane, select the Fast Backward variable selection method. Keep the significance level at .01. I. Create the logistic model. On the menu bar, click : (Menu) = Interface options = Enable auto-refresh. e Examine the Fit Summary panel. How many of the 23 input variables are not included in this model? ¢ In this model, are any of the insignificant variables that are not included classification effects? e What is the value of the AIC statistic? 2. Examining Additional Logistic Regression Results a. Open the details table and click the Parameter Estimates tab. 1) Click the Estimate column heading twice to sort the parameter estimates. Then determine which parameter had the largest estimate. What was the value? 2) Click the Response Profile tab to determine how many of these customers made donations. b. Close the details table. Copyright © 2020, SAS Institute Inc., Cary, North Carolina, USA. ALL RIGHTS RESERVED.
4.1 Logistic Regression 4-45 ¢. Maximize the assessment plot to gain access to the assessment charts. 1) Examine the lift chart to determine the advantages of using this model for prediction. How does this model compare to the Best model? 2) Select the ROC chart and report the KS statistic and the cutoff value. 3) What is the prediction cutoff value that is used in the current logistic regression model? 4) Select the misclassification chart to determine whether this model predicts more true positives or more true negatives. d. Save your report as Practice 6. End of Practices Copyright © 2020, SAS Institute Inc., Cary, North Carolina, USA. ALL RIGHTS RESERVED.
4.2 Modeling with Group-By Variables 4-53 @ Practice 'S In this practice, you continue to use the PVA data set to investigate segmenting a logistic regression model interactively by gender. 3. Adding an Interactive Group-By Variable in SAS Visual Statistics a. Return to your remote desktop client machine. If your session timed out, sign in. Use Student as the user ID and Metadata0 as the password. b. Open your saved report, Practice 6, from the previous practice. ¢. On the Roles option, move the variable Gender from the Classification effects role to the Group by role. Do this by dragging and dropping the variable or by right-clicking, removing, and re-adding. d. Examine the Fit Summary pane to discover that there are several observations with a value of U. Create a report filter to remove this value from the current model. e. Maximize the Fit Summary window. Is Age a significant variable for both the Male and Female segmented models? Is Promotion Count Card 36 Months 36 Months significant for both segmented models? f. Remove Gender from the model completely. g. On the Roles pane, move the variable Status Category 96NK from the Classification effects role to the Group by role. e Which effects are important to every BY-group level? e What is the AICC statistic for the N BY-group level? h. Open the details table and click the Type Ill Test tab. e How many effects are significant at .01 for the N BY-group level? i. Save your report as Practice 6_Groupby. End of Practices Copyright © 2020, SAS Institute Inc., Cary, North Carolina, USA. ALL RIGHTS RESERVED.
Your preview ends here
Eager to read complete document? Join bartleby learn and gain access to the full version
- Access to all documents
- Unlimited textbook solutions
- 24/7 expert homework help