ROC curve |
- Sensitivity vs
 - specificity
|
- True positive rate vs false positive rate
|
- A random guess algorithm is a
 line
|
- Area under the curve measures accuracy (or "discrimination")
|
- A perfect algorithm has an area of
|
- A random (or useless) algorithm has an area of
|
|
- 0.9-1.0: excellent; 0.8-0.9: good; 0.7-0.8: fair; 0.6-0.7: poor; 0.5-0.6:
fail
|
