1- Analyze and determine based on the 3 performance Vectors below Entropy, logistic regression, and Optimize Parameter Attrition why is this company is losing customers (attrition rate) by using the primary attribute of attrition rate to identify patterns/reasons that people are leaving ( analyze the data results and identify anomalies or reasons to believe people are leaving the company (lack of benefits, high APR, age groups, income level, male and female, etc.)

2-Which model is best is beast to use for a business proposal?

1- Entropy - Attrition:

PerformanceVector

PerformanceVector:

accuracy: 94.11%

ConfusionMatrix:

True: 0 1

0: 2494 102

1: 77 365

AUC: 0.909 (positive class: 1)

precision: 82.58% (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 2494 102

1: 77 365

false_positive: 77.000 (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 2494 102

1: 77 365

false_negative: 102.000 (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 2494 102

1: 77 365

true_positive: 365.000 (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 2494 102

1: 77 365

true_negative: 2494.000 (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 2494 102

1: 77 365

sensitivity: 78.16% (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 2494 102

1: 77 365

specificity: 97.01% (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 2494 102

1: 77 365

2- Logistic Regression - Attrition

PerformanceVector

PerformanceVector:

accuracy: 70.10%

ConfusionMatrix:

True: 0 1

0: 1167 16

1: 619 322

AUC: 0.926 (positive class: 1)

precision: 34.22% (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 1167 16

1: 619 322

false_positive: 619.000 (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 1167 16

1: 619 322

false_negative: 16.000 (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 1167 16

1: 619 322

true_positive: 322.000 (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 1167 16

1: 619 322

true_negative: 1167.000 (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 1167 16

1: 619 322

sensitivity: 95.27% (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 1167 16

1: 619 322

specificity: 65.34% (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 1167 16

1: 619 322

3- Optimize Parameter - Attrition

PerformanceVector

PerformanceVector:

accuracy: 92.75%

ConfusionMatrix:

True: 0 1

0: 1696 51

1: 103 274

AUC: 0.949 (positive class: 1)

precision: 72.68% (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 1696 51

1: 103 274

true_positive: 274.000 (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 1696 51

1: 103 274

true_negative: 1696.000 (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 1696 51

1: 103 274

sensitivity: 84.31% (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 1696 51

1: 103 274

specificity: 94.27% (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 1696 51

1: 103 274

positive_predictive_value: 72.68% (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 1696 51

1: 103 274

negative_predictive_value: 97.08% (positive class: 1)

ConfusionMatrix:

True: 0 1

0: 1696 51

1: 103 274