chartl
|
5a79f16ea4
|
Fixed an edge case where an exception was thrown if either of the sets was empty for the MWU test. Also altered the output format so U itself is not printed (which though interesting, isn't so useful for recalibration), but rather a value I call V (really the deviation of U from its expectation).
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5490 348d0f76-0448-11de-a6fe-93d51630548a
|
2011-03-22 16:28:44 +00:00 |
|
chartl
|
60ddc08cdf
|
Added a boatload of new case-control association modules. Switched the U-test to use longs rather than ints (it just so happened that I overflowed and started getting negative U statistics. Not good.) Added the ALL association type for ease of specifying that we want to throw the book at something. Added an svn-commit.tmp~ because i can't get rid of it even with --force. Hopefully I can remove it after.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5386 348d0f76-0448-11de-a6fe-93d51630548a
|
2011-03-06 21:58:12 +00:00 |
|
chartl
|
a40a8006b5
|
Added in unit tests for the statistics calculated by the test runner; and bug-fixes to the calculations; so we have some assurance that the statistics coming out the back-end are correct.
git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5380 348d0f76-0448-11de-a6fe-93d51630548a
|
2011-03-06 16:54:02 +00:00 |