Pre-Loss Analysis to determine a risk
- Could be used to determine credit risk, effectiveness of a promotion
- Segment population into informative attributes
- Sort a target variable by informative characteristics
Predictive Modeling
Example
- Target Variable - Sale (Purchases good = Triangle, Does not Purchase = Circle)
- Treatment - Promotion (Customer Receives a Promo = Green, Does Not Receive Promo = Red)
Overall
| shape | count |
|---|---|
| circle | 5 |
| triangle | 5 |
Segent by Color = Green
Green Table
| shape | count |
|---|---|
| circle | 2 |
| triangle | 3 |
Segment Color = Red
Red Table
| shape | count |
|---|---|
| circle | 3 |
| triangle | 2 |
Supervised Segmentation
Segment By Color: Overall
| shape | count |
|---|---|
| circle | 50 |
| triangle | 50 |
Segment by Color: Green
| shape | count |
|---|---|
| circle | 30 |
| triangle | 10 |
Segment by Color: Red
| shape | count |
|---|---|
| circle | 20 |
| triangle | 40 |