Opened 3 years ago
Last modified 2 years ago
#3129 reviewing defect
VariableRanges of RegressionProblemData are invalid if one NaN is present
| Reported by: | mkommend | Owned by: | chaider |
|---|---|---|---|
| Priority: | medium | Milestone: | HeuristicLab 3.3.17 |
| Component: | Problems.DataAnalysis | Version: | trunk |
| Keywords: | Cc: |
Description
If any value of a dataset variable is missing (NaN) the whole calculated interval of the variable (variable range) is set to NaN. The reason therefore is the reuse of intervals that have been implemented for a different purpose. Hence, the values of a variable should be filtered for NaNs before the interval / range of a variable is calculated.
Change History (5)
comment:1 Changed 3 years ago by mkommend
- Status changed from new to accepted
comment:2 Changed 3 years ago by mkommend
comment:3 Changed 3 years ago by mkommend
- Owner changed from mkommend to chaider
- Status changed from accepted to reviewing
@chaider please check the changes in r17999 and then forward the ticket to gkronber for review.
comment:4 Changed 3 years ago by mkommend
r18000: Fix compilation error in RegressionProblemData.
comment:5 Changed 2 years ago by gkronber
Closed on github.
Note: See
TracTickets for help on using
tickets.
r17999: Work on variable range calculation of datasets.
Renamed method to calculate variable ranges to a more appropriate method name. Added possibility to ignore NaNs for range calculation.