Scribd
Upload
36 views8 pages

Answers Apate264 EXAM2

The document analyzes customer churn based on various factors such as complaint codes and average days delinquent, providing metrics like accuracy, sensitivity, and predictive values. It presents a comparison of outcomes when using a predictive model versus not using one, detailing customer responses and financial gains or losses. The results indicate a significant net gain when the model is applied compared to traditional methods.

Uploaded by

apate264
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as XLSX, PDF, TXT or read online on Scribd
36 views8 pages

Answers Apate264 EXAM2

The document analyzes customer churn based on various factors such as complaint codes and average days delinquent, providing metrics like accuracy, sensitivity, and predictive values. It presents a comparison of outcomes when using a predictive model versus not using one, detailing customer responses and financial gains or losses. The results indicate a significant net gain when the model is applied compared to traditional methods.

Uploaded by

apate264
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as XLSX, PDF, TXT or read online on Scribd
You are on page 1/ 8

Avg_Days_Delinquent

Complaint Code = Call Quality


Complaint Code = Inaccurate Sales Inf
Precent_Increase_MOM
For one unit increased on Avg_Days_Delinquent, the odds of a Customer Churning Increase
For one unit increased on Complaint Code = Call Quality, the odds of a Customer Churning
For one unit increased on Complaint Code = Inaccurate Sales Inf, the odds of a Customer C
For one unit increased on Precent_Increase_MOM, the odds of a Customer Churning Decrea
Matrix Predicted Matrix
Actual 0 1 Actual
0 True Negative (TN) False Positive (FP) 0
1 False Negative (FN) True Positive (TP) 1

Formulas
Error Rate or Misclassification Rate Matrix
FalsePositive + FalseNegative Actual 0
All Cases (TN + TP + FN + TN) Actual 1

Accuracy Matrix
TruePositive + TrueNegative Actual 0
All Cases (TN + TP + FN + TN) Actual 1

Specificity Matrix
True Negative Actual 0
(True Negative + False Positive) Actual 1

Sensitivity Matrix
True Positive Actual 0
(True Positive + False Negative) Actual 1

Positive predictive Value (PPV) Matrix


True Positive Actual 0
(True Positive + False Positive) Actual 1

Negative Predictive Value (NPV) Matrix


True Negative Actual 0
(True Negative + False Negative) Actual 1
Predicted
0 1
656 152
173 590

Predicted 0 Predicted 1 Error Rate or Misclassification Rate


True Negative (TN) False Positive (FP) 21%
False Negative (FN) True Positive (TP)

Predicted 0 Predicted 1 Accuracy


True Negative (TN) False Positive (FP) 79%
False Negative (FN) True Positive (TP)

Predicted 0 Predicted 1 Specificity


True Negative (TN) False Positive (FP) 81%
False Negative (FN) True Positive (TP)

Predicted 0 Predicted 1 Sensitivity


True Negative (TN) False Positive (FP) 77%
False Negative (FN) True Positive (TP)

Predicted 0 Predicted 1 Positive predictive Value (PPV)


True Negative (TN) False Positive (FP) 80%
False Negative (FN) True Positive (TP)

Predicted 0 Predicted 1 Negative Predictive Value (NPV)


True Negative (TN) False Positive (FP) 79%
False Negative (FN) True Positive (TP)
r Rate or Misclassification Rate
error in the data

Accuracy

Specificity

Sensitivity

sitive predictive Value (PPV)

gative Predictive Value (NPV)


Question C
New Sample 15,000
Contact Budget $ 225,000
Cost of contact $ 60
Purchase
Model Cost

Max of ppl contact 3750

W/model Wo/Model
Depth 25% 100%

Cumulative % of Response 88.2952% 48.5678%

# Customers 3311 1821

Purchase $ - $ -

Gain/Loss $ (225,000) $ (225,000)

Question D
New Sample 15,000
Contact Budget $ 225,000
Cost of contact $ 60
Purchase $ 400
Model Cost $ 25,000

Max of ppl contact 3750

W/model Wo/Model
Depth 25% 100%

Cumulative % of Response 88.2952% 48.5678%

# Customers 3311 1821

Purchase $ 1,324,400 $ 728,400

Gain/Loss $ 1,074,400 $ 503,400


QUETION
If you use your model and contact the maximum number of customers your budget allows
If you do not use the model and contact the maximum number of customers your budget a
how much net gain/loss will the company make if the model is used?
How much net gain/loss will the company make if the model is not used?
ANSWER
3311 Customers
1821 Customers
$ 1,074,400
$ 503,400

You might also like