Open HeuristicLab-Trac-Bot opened 14 years ago
Created branch for population diversity analysis for symbolic regression in r4877.
Added first version of variables usage population diversity analyzer in r4881.
Added abstract base class for population diversity analyzers for symbolic regression and simplified management of population diversity analyzers in symbolic regression problem base in r4885.
Started working on fine grained structural similarity estimation for structural population diversity in GP in r4886.
Worked on retrieval of genetic items in r4917. Idea: Collect all descendants' lists recursively and then add new items using the retrieved ones; this should save lots of time and reduce complexity of the items retrieval process (because we don't need references to parents of nodes etc.).
Worked on structural population diversity analysis in r4928:
- Improved definition of genetic items.
- Implemented recursive collection of all genetic items of symbolic expression trees.
- Added FineGrainedStructuralPopulationDiversityAnalyzer to SymbolicRegressionProblemBase.
Worked on structural population diversity analysis in r4934:
- Added storing of descendant nodes.
- Added storing of level information.
- Added level filtering.
- Added prevention of multiple consideration of descendant nodes.
Improved and pruned similarity function for genetic information items in r4938.
Implemented calculation of structural similarity of symbolic expression trees in r4941.
Added parameters for similarity calculations in r4942.
Added parameters for similarity calculations in r4948.
Implemented management of genetic information items in dictionaries in r4949.
Branched DataAnalysis plugin into DataAnalysis.PopulationDiversity branch in r4951.
Open issues:
- Population diversity analyzers should be moved to DataAnalysis plugin.
- Interfaces also have to be modified accordingly.
- Population diversity analyzers should be added in regression problem and in classification problem.
Moved GeneticInformationItem class into separate file in r5024.
Stephan: what is the state of this ticket? Do you think it's feasible to move these changes to the trunk soon?
This topic is still pending and on my list for the next months. I will synchronize with Bogdan and check in how far his developments can be combined with the here developed fine grained population diversity.
Please rename the branch by prepending it with the ticket number
Issue migrated from trac ticket # 1278
milestone: HeuristicLab 3.3.x Backlog | component: Problems.DataAnalysis | priority: medium
2010-11-17 21:38:49: @StephanWinklerHagenberg created the issue