Differences

This shows you the differences between two versions of the page.

--- kvalobs:kvoss:system:qc2:test:algorithms:diptest [2011-05-08 10:10:35]
paule
+++ kvalobs:kvoss:system:qc2:test:algorithms:diptest [2022-05-31 09:29:32] (current)
@@ Line 3: / Line 3: @@
 Output of various tests. The latest is always posted at the top of the page.
-{{:kvoss:system:qc2:requirements:algorithms:dip.png|}}
+{{kvalobs:kvoss:system:qc2:requirements:algorithms:dip.png|}}
    * Also see the application of [[https://kvalobs.wiki.met.no/doku.php?id=kvoss:system:qc2:test:algorithms:akima#using_akima_in_the_diptest_algorithm|Akima to the Diptest]]
@@ Line 9: / Line 9: @@
 ====== Simulated Test Data after Flag 9.7 Updates ======
-*Review*
+**Review**
-The lats results of testing this algorithm are provided below (subsections labeled "Results"). These outcomes were reviewed April 27-28 at met.no, the following conclusions were reached, that may be considered as part of the overall recommendation, and related conditions, to now put the algorithm into operations.
+The last results of testing this algorithm are provided below (subsections labeled "Results"). These outcomes were reviewed April 27-28 at met.no. The following conclusions were reached, that may be considered as part of the overall recommendation, and related conditions, to now put the algorithm into operations.
    * The results of the DipTest with regards the flags set at the final step and the interpolation performed are in general OK  but the applicability of the DipTest to all of the parameters originally proposed should be checked.
-   * Review of the results sets suggests that paramid=104 should not be checked and should be removed from the paramid listing. (Difficult to correctly handle the periods where the bucket is manually emptied in the time series).
+   * For example, review of the results sets suggests that paramid=104 should not be checked and should be removed from the paramid listing. (Difficult to correctly handle the periods where the bucket is manually emptied in the time series).
-   * The
+   * For some examples the flagging apears not to be correctly set prior to the DipTest (e.g. useinfo not set according to teh controlinfo), but in the end the result after the DipTest is OK. This may be a bug in a QC1 control or just because data from 2008 (rewritten here to be 2018) was used for the test.
+**Results**
-*Results*
 The following data was taken from 2008 and then modified to create and emphasise some dips. Controlflags may also have been modified to ensure that the DipTest ran - this simulation is just to get some more DipTest results to look at, but simulating data may result in impossible flag settings etc ... so please be wary.