Blame view

tools/sctk-2.4.10/src/hubscr/test1.base/Ensemble.stats.unified 3.64 KB
8dcb6dfcb   Yannick Estève   first commit
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
  ,------------------------------------------------------------------------------.
  |                  Composite Report of All Significance Tests                  |
  |                                For the  Test                                 |
  |                                                                              |
  |                            Test Name                            Abbrev.      |
  |      ------------------------------------------------------     -------      |
  |            Matched Pair Sentence Segment (Word Error)             MP         |
  |      Signed Paired Comparison (Speaker Word Error Rate (%))       SI         |
  |        Wilcoxon Signed Rank (Speaker Word Error Rate (%))         WI         |
  |                     McNemar (Sentence Error)                      MN         |
  |                                                                              |
  |                                                                              |
  |------------------------------------------------------------------------------|
  |   Test    ||                | lvc_hyp.ctm | lvc_hyp2.ctm  ||      Test       |
  |  Abbrev.  ||                |             |               ||     Abbrev.     |
  |-----------++----------------+-------------+---------------++-----------------|
  |    MP     ||  lvc_hyp.ctm   |             |  ~   1.000    ||       MP        |
  |    SI     ||                |             |  ~   1.000    ||       SI        |
  |    WI     ||                |             |  ~   1.000    ||       WI        |
  |    MN     ||                |             |  ~   1.000    ||       MN        |
  |-----------++----------------+-------------+---------------++-----------------|
  |    MP     ||  lvc_hyp2.ctm  |             |               ||       MP        |
  |    SI     ||                |             |               ||       SI        |
  |    WI     ||                |             |               ||       WI        |
  |    MN     ||                |             |               ||       MN        |
  |------------------------------------------------------------------------------|
  |  These significance tests are all two-tailed tests with the null hypothesis  |
  |  that there is no performance difference between the two systems.            |
  |                                                                              |
  |  The first column indicates if the test finds a significant difference       |
  |  at the level of p=0.05.  It consists of '~' if no difference is             |
  |  found at this significance level.  If a difference at this level is         |
  |  found, this column indicates the system with the higher value on the        |
  |  performance statistic utilized by the particular test.                      |
  |                                                                              |
  |  The second column specifies the minimum value of p for which the test       |
  |  finds a significant difference at the level of p.                           |
  |                                                                              |
  |  The third column indicates if the test finds a significant difference       |
  |  at the level of p=0.001 ("***"), at the level of p=0.01, but not            |
  |  p=0.001 ("**"), or at the level of p=0.05, but not p=0.01 ("*").            |
  |                                                                              |
  |  A test finds significance at level p if, assuming the null hypothesis,      |
  |  the probability of the test statistic having a value at least as            |
  |  extreme as that actually found, is no more than p.                          |
  `------------------------------------------------------------------------------'