#
source:
orange/docs/widgets/rst/evaluate/liftcurve.rst
@
11359:8d54e79aa135

Revision 11359:8d54e79aa135, 2.2 KB checked in by Ales Erjavec <ales.erjavec@…>, 14 months ago (diff) |
---|

Line | |
---|---|

1 | .. _Lift Curve: |

2 | |

3 | Lift Curve |

4 | ========== |

5 | |

6 | .. image:: ../icons/LiftCurve.png |

7 | |

8 | Shows the lift curves and analyzes them. |

9 | |

10 | Signals |

11 | ------- |

12 | |

13 | Inputs: |

14 | - Evaluation Results (orngTest.ExperimentResults) |

15 | Results of classifiers' tests on data |

16 | |

17 | |

18 | Outputs: |

19 | - None |

20 | |

21 | Description |

22 | ----------- |

23 | |

24 | Lift curves show the relation between the number of instances which were |

25 | predicted positive and those of them that are indeed positive. This type of |

26 | curve is often used in segmenting the population, e.g., plotting the number |

27 | of responding customers against the number of all customers contacted. Given |

28 | the costs of false positives and false negatives, it can also determine the |

29 | optimal classifier and threshold. |

30 | |

31 | .. image:: images/LiftCurve.png |

32 | |

33 | Option :obj:`Target class` chooses the positive class. In case there are |

34 | more than two classes, the widget considers all other classes as a single, |

35 | negative class. |

36 | |

37 | If the test results contain more than one classifier, the user can choose |

38 | which curves she or he wants to see plotted. :obj:`Show convex lift hull` |

39 | plots a convex hull over lift curves for all classifiers. The curve thus |

40 | shows the optimal classifier (or combination thereof) for each desired TP/P |

41 | rate. The diagonal line represents the behaviour of a random classifier. |

42 | |

43 | The user can specify the cost of false positives and false negatives, and |

44 | the prior target class probability. :obj:`Compute from Data` sets it to the |

45 | proportion of examples of this class in the data. The black line in the |

46 | graph, which corresponds to the right-hand axis, gives the total cost for |

47 | each P ration for the optimal classifier among those selected in the list |

48 | box on the left. The minimum is labelled by the optimal classifier at that |

49 | point and the related cost. |

50 | |

51 | The widget allows setting costs from 1 to 1000. The units are not important, |

52 | as are not the magnitudes. What matters is the relation between the two |

53 | costs, so setting them to 100 and 200 will give the same result as 400 and 800. |

54 | |

55 | Example |

56 | ------- |

57 | |

58 | At the moment, the only widget which give the right type of the signal |

59 | needed by the Lift Curve is :ref:`Test Learners`. The Lift Curve will hence |

60 | always follow Test Learners and, since it has no outputs, no other widgets |

61 | follow it. Here is a typical example. |

62 | |

63 | .. image:: images/ROCLiftCalibration-Schema.png |

**Note:**See TracBrowser for help on using the repository browser.