You are here:
Vanderbilt Biostatistics Wiki
>
Main Web
>
Wfccm2Testing
(14 Dec 2005,
JoanZhang
)
(raw view)
E
dit
A
ttach
%RED% red content are potential problems are found %ENDCOLOR% * Purpose : testing WFCCM2 regularly to ensure that problems are found and dealt with. * Methods * functional testing: test particular functions or code modules * integration testing: testing functionallity of combined parts (overall) of WFCCM * Data Set (working folder research drive \WFCCM2_Testing ) * black data set * a data set with pure random values * a data set that with extreme values * white data set (data sets that have been analyzed) * an AUC data set * a Zero data set * Tasks * priority : visualization (functionallity, state files ....) * others ---+++ Testing report * report 1 (Nov. 1): (per an Investigator's request) visualzation is able to display an averaged data set with more than 2 groups according to their unique column id. However, we may need make it be able to load groups with duplicate columns and use same color to display each group. * ex: in average data, we have three columns belong to three groups, let's say stage1 stage2, and stage3. If we have total 5 columns in the average data set, 2 of them are stage1, 2 of them are stage2, we should able to display 5 curves but only have three colors. * status: done ---++ Testing Checklist (based on VS 2005 and .NET 2.0) * %Y% data loading : logged and unlogged data sets (AUC, Zero, Microarray(with missing)) * %Y% data grouping : regular 2 groups, irreglar groups * %Y% score generation: result,load score, add score, save a score * criterial: add%Y%, load%Y%, remove%Y%, STD function%Y%, PreFilter, percentage for data * distance run : result, progress bar, topN and cutoff * %Y% cluster * %Y% ballgraph * %Y% ROC curve * visualzation * all buttons * winner panel * %Y% box plot * %Y% Option label%Y% * %Y% state file * %Y% quality control * corss validation * overall performance ---+++ problems found %RED% * it can not open to view some score files through application if there are too many features/genes in the data set, however scores are able to be generated and saved , and could be viewed in local (%GREEN%fixed %Y%%ENDCOLOR%) * ROC * no ROC generated when cutoff with mutiple numPass * ROC curve can not be displayed (%GREEN%fixed%ENDCOLOR% %Y%) * Do we need change the way of ROC to show up to make it consistent with cluster and ballgraph? use click prediction model instead of going "view"? (%GREEN%fixed%ENDCOLOR% %Y%) * error occurs when launching ROC with 1 winner (%GREEN%fixed %Y% %ENDCOLOR%) * does testing set need ROC also? * Cluster * cluster node can not be removed from tree (%GREEN%fixed%ENDCOLOR% %Y%) * exception occurs when trying generate with top1 or top2 run * gene tree should move to the left side, array tree should move to the top; or at least the way of gene tree lable need to be change, therefore, these lables is still readble after rotation. * logical of score generation of log10 data set ??? (need ask Will) %ENDCOLOR% ---+++ Testing folder
E
dit
|
A
ttach
|
P
rint version
|
H
istory
: r20
<
r19
<
r18
<
r17
|
B
acklinks
|
V
iew topic
|
Edit
w
iki text
|
M
ore topic actions
Topic revision: r20 - 14 Dec 2005,
JoanZhang
Main
Department Home Page
Biostatistics Graduate Program
Vanderbilt University Medical Center
Main Web
Main Web Home
Search
Recent Changes
Changes
Topic list
Biostatistics Webs
Archive
Main
Sandbox
System
Register
|
Log In
Copyright © 2013-2022 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki?
Send feedback