3 - Answer Key.pdf-VA & VS Software - Ho...
3_-_Answer_Key.pdf-VA & VS Software - Homework
Showing 7-8 out of 17
3 - Answer Key.pdf-VA & VS Software - Homework
3_-_Answer_Key.pdf-VA & VS Software - Homework
3 - Answer Key.pdf-VA & VS Software...
3_-_Answer_Key.pdf-VA & VS Software - Homework
Page 7
(Results may vary, but the basic premise should be similar to what I wrote above)
ll.
Select each of the other cluster IDs individually in the parallel coordinates plot to examine the
characteristics of each cluster.
mm.
Show the summary table.
1)
Which cluster has the smallest Within-Cluster SS?
Answer: Cluster 4
2)
Do we want to maximize or minimize the Within-Cluster SS for K-Means Clustering?
Answer: Minimize
nn.
Hide the summary table, minimize the Parallel Coordinates plot, and re-focus on the Cluster
Matrix.
oo.
Right-click any of the cells in the cluster matrix. Select
Derive a Cluster ID Variable
. A new
variable is created and appears in the Data pane as
Cluster ID (1)
.
pp.
Make
Decision Tree
(the model from the previous exercise) the active model in your analysis
window and assign
Cluster ID 1
as an additional predictor variable.
1)
Which variable does the tree initially split on now?
Answer: Cluster ID (1)
2)
How many leaf nodes does the new decision tree have compared to the original decision
tree, which was 15?
Answer: 7
qq.
Take a quick look at your assessment plots for your updated Decision Tree analysis. Hopefully
you see some positive updates…
rr.
Minimize the decision tree analysis, and make
Logistic – with IM and VarSel
the active model
and assign
Cluster ID 1
as an additional classification effect.
1)
What is the R square for the model compared to the original model, which was 0.2258?
Answer: .5444
ss.
Make
Logistic
the active model. Assign
Cluster ID 1
as a classification effect.
tt.
Select
Model Comparison
on the toolbar and compare the three models. Specify Level
1
in the
Model Comparison window. Select all models and click
OK
.
If it appears, close the
Number of observations for all models do not match
message
window.
uu.
Adjust the Prediction Cuttoff rate to 0.28
1)
Which model is selected when the fit statistic is misclassification?
Answer: Logistic - with IM and VarSel
2)
Which model is the next best model based on misclassification?
Answer: Logistic
vv.
Save the exploration as
VA Software HW 2
in the
My Folder
directory
.
ww.
Make
Logistic – with IM and VarSel
the active model, select Loyalty Status from the
classification effect role, and drag it down to the Group By role.


Page 8
1)
At a high level, what just happened? Please include screenshot with your answer.
Answer: Four logistic regression models were built, one for each level of the variable
Loyalty Status.
Image:
xx.
Are you having fun yet?
yy.
Save the exploration as
VA Software HW 2
in the
My Folder
directory
.


Ace your assessments! Get Better Grades
Browse thousands of Study Materials & Solutions from your Favorite Schools
George Washington Univers...
George_Washington_University
School:
Visualization_for_Analytics
Course: