Assignment 7A.4
pptx
keyboard_arrow_up
School
University of Louisiana, Lafayette *
*We aren’t endorsed by this school
Course
510
Subject
Industrial Engineering
Date
Dec 6, 2023
Type
pptx
Pages
7
Uploaded by SuperHumanPencilOyster3
Assignment 7A.4
Comparing Classification Models
Start with the workflow you used for Assignment 6B.3
(Decision Tree) with the Default Data
1) Here's what 6B.3 should look like, be sure you
have unchecked the Default_Indicator column in
the Excel Reader node, as well as set up the
Partitioning node with the usual settings - 70%
relative, draw randomly, with a seed of 1234
2)
Add a Binary Classification
Inspector node as follows:
(use
the same settings for the
Decision Tree Predictor and
Binary Classification Inspector
Nodes as provided on Slide #5)
RUN the workflow and review the results of the Binary Classification
Inspector.
Paste a screenshot of the output (including the Confusion
Matrix) which includes your name below.
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
Now, open the workflow you used for Assignment
6A.3 (Logistic Regression) with the Default Data
Here's what 6A.3 should look like
Your 3
rd
model in that workflow should have
been set up as follows: It uses all 3 variables,
just as we did in the decision tree model
Now, open the workflow you used for Assignment
6A.3 (Logistic Regression) with the Default Data
The Logistic Regression Predictor Node should be
set up as follows:
With the Binary Classification Inspector
settings as follows:
RUN the workflow for the
3
rd
Logistic Regression Model
and review the
results of the Binary Classification Inspector.
Paste a screenshot of the
output (including the Confusion Matrix) which includes your name below.
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
Answer the following questions:
Using the AUC metric, which model would you select as the best model (Decision Tree or Logistic Regression)
and why?
1.
Logistic Regression is what I would select as the best model because the AUC and Accuracy are closer
to 1.
Based on the metric, sensitivity, which model (Decision Tree or Logistic Regression) is best and why?
2.
Based on the metric, sensitivity, the Decision Tree model is best because the sensitivity metric is
higher than that of the Logistic Regression model, meaning fewer false negatives exist.
Which is of greater concern for the bank, false negatives or false positives? Explain your reasoning.
3.
False negatives are of greater concern because a false negative indicates that there is no issues with a
transaction when in actuality, an issue does exist.
Given your answer in the question above, how can the threshold be adjusted (increased or decreased) to
improve the results of the model for the Bank?
Decreasing the threshold increases the sensitivity, decreasing the amount of false
negatives.