Free cookie consent management tool by TermsFeed Policy Generator

Opened 14 years ago

Closed 13 years ago

Last modified 13 years ago

#1011 closed feature request (done)

RunCollection view for variable impact analysis

Reported by: gkronber Owned by: mkommend
Priority: medium Milestone: HeuristicLab 3.3.1
Component: Problems.DataAnalysis Version: 3.3.1
Keywords: Cc:

Description


Change History (19)

comment:1 Changed 14 years ago by gkronber

  • Version changed from 3.4 to 3.3

comment:2 Changed 14 years ago by gkronber

  • Owner changed from gkronber to mkommend

comment:3 Changed 14 years ago by mkommend

implemented first version of RunCollectionVariableImpactView r3969.

comment:4 Changed 14 years ago by mkommend

added p-values for variable impacts r3975.

comment:5 Changed 14 years ago by mkommend

  • Status changed from new to assigned

comment:6 Changed 14 years ago by mkommend

fixed RunCollectionVariableImpactView r4010

comment:7 Changed 14 years ago by mkommend

corrected namespace of RunCollectionVariableImpactView and changed test direction r4037

comment:8 Changed 14 years ago by gkronber

The variable impact view should be improved.

  • Variables which are not used in any runs should not be shown
  • The significance test against mean 0 should be changed to a test if the location of the variable impact distribution of the current variable is significantly different to the location of the variable impact distribution of an unimportant variable. (The variable with the smallest average impact can be used as a reference.
  • The students t-test shouldn't be used because it is clearly not applicable (the variable impact values are not normally distributed). Instead the non-parametric Mann-Whitney-U should be used to compare the location of variables x_i and x_reference.
  • Only two-tailed tests should be used.
  • The median rank of the variable should be shown (as this is the value that is tested by Mann-Whitney-U.

comment:9 Changed 14 years ago by gkronber

  • Owner changed from mkommend to gkronber
  • Status changed from assigned to new
  • Version changed from 3.3 to 3.3.1

comment:10 Changed 14 years ago by gkronber

  • Status changed from new to assigned

comment:11 Changed 14 years ago by gkronber

Improved variable impact view with r4124

comment:12 Changed 14 years ago by gkronber

Made variable frequency analyzer more efficient and removed subtraction of base line (variable impacts are now in the range 0..1 instead of -1 .. 1) with r4125.

comment:13 Changed 14 years ago by gkronber

r4143:

  • Fixed calculation of p-values for variable impacts (the calculation procedure was implemented incorrectly because in r4121).

comment:14 Changed 13 years ago by gkronber

  • Status changed from assigned to accepted

comment:15 Changed 13 years ago by gkronber

  • Owner changed from gkronber to mkommend
  • Status changed from accepted to reviewing

comment:16 Changed 13 years ago by mkommend

  • Status changed from reviewing to readytorelease

comment:17 Changed 13 years ago by mkommend

  • Resolution set to done
  • Status changed from readytorelease to closed

comment:18 Changed 13 years ago by swagner

  • Milestone changed from Iteration 4 to Current

Milestone Iteration 4 deleted

comment:11 Changed 13 years ago by swagner

  • Milestone changed from Current to HeuristicLab 3.3.0

Milestone Current deleted

Note: See TracTickets for help on using tickets.