milestone: HeuristicLab 3.3.x Backlog | component: Problems.DataAnalysis.Symbolic.Views | priority: medium
2016-07-04 19:15:36: @gkronber created the issue
It is possible to drop ProblemData objects onto regression solutions to evaluate a particular solution on a new dataset. In this case the dataset is changed so that the test partition contains all observations (there is no training partition).
However, the simplifier view calculates node impacts based on training samples only and since there are no training samples in this case an exception occurs.
(It is also not possible to optimize parameters using the structure of the current solution to fit the parameters to the new dataset)
2016-09-17 10:05:20: @gkronber changed title from Exception when dropping a new problem data onto the simplified view for regression solutions to Exception when dropping a new problem data onto the simplifier view for regression solutions
Issue migrated from trac ticket # 2623
milestone: HeuristicLab 3.3.x Backlog | component: Problems.DataAnalysis.Symbolic.Views | priority: medium
2016-07-04 19:15:36: @gkronber created the issue