Improve structure of data analysis specific classes

[gkronber]

Created branch for refactoring with r5471.

[mkommend]

Added SymbolicExpressionTreeEncoding, -Views, ArtificialAnt and -Views 3.4 plugins with r5473.

[gkronber]

r5474: Created empty projects for data analysis problem classes (version 3.4)

[mkommend]

Added Algorithm.DataAnalysis 3.4 with r5475.

[mkommend]

Added DataAnalysis.Views 3.4 with r5476.

[gkronber]

r5477: added projects to solution file, updated files.txt and svn:ignore properties.

[mkommend]

Prepared 3.4 project stubs for refactoring and added interfaces in Problems.DataAnalysis 3.4 with r5484.

[mkommend]

Updated interfaces of DataAnalysis 3.4 with r5486.

[mkommend]

Added interfaces for Classification 3.4 with r5487.

[mkommend]

Added interfaces for Regression 3.4 with r5488.

[gkronber]

r5490: minor adoptions in generic interfaces for data analysis

[gkronber]

r5491: updated project file.

[mkommend]

r5492: Refactored Encodings.SymbolicExpressionTree.

[mkommend]

r5495 - Created interfaces in DataAnalysis.Symbolic.

[mkommend]

r5496 - Merged DataAnalysis.Regression and Classification into DataAnalysis.

[mkommend]

r5497 - Removed remaining files from deleted projects.

[mkommend]

r5498 - Added Single- and MultiObjectiveSymbolicDataAnalysisEvaluators.

[gkronber]

r5499: changes in symbolic expression tree encoding.

[mkommend]

r5500: Added all single objective symbolic regression analyzers.

[mkommend]

r5501: Corrected namespaces and added symbolic classification specific evaluators.

[mkommend]

r5505: worked on single and multi objective evaluators.

[mkommend]

r5506: removed unnecessary directory.

[mkommend]

r5507: Commited remaining changes regarding OnlineEvaluators and solutions.

[mkommend]

r5508: corrected IDataAnalysisSolution.

[mkommend]

r5509: Made evaluators and problems generic to create the parameters correctly.

[gkronber]

r5510: Fixed compiler errors in symbolic expression tree encoding

[mkommend]

r5513: Adapted views to new symbolic expression tree encoding.

[mkommend]

r5514: Fixed compilation errors in Problems.DataAnalysis.*.

[mkommend]

r5517: Fixed compilation errors in Problems.ArtificalAnt.

[mkommend]

r5519: Fixed all compilation errors and reformatted some files.

[mkommend]

r5521: Corrected wiring of artificial ant problem.

[mkommend]

r5528: Updated ArtificialAntProblem.

[gkronber]

r5529: Removed grammar specific members from ISymbolicExpressionTreeNode interface.

[gkronber]

r5530: added static method to calculate mean and variance of a double enumerable.

[gkronber]

r5532: copied symbol classes into Problems.DataAnalysis.Symbolic plugin.

[gkronber]

r5540: added implementations of DataAnalysisProblem and RegressionProblem

[gkronber]

r5542: improved event propagation in DataAnalysisProblemData and RegressionProblemData

[gkronber]

r5547: moved range limitation of estimated values to upper and lower limits into HeuristicLab.Common plugin.

[gkronber]

r5549: unified size/height vs. length/depth terminology for symbolic expression trees and adapted unit tests for symbolic expression tree encoding version 3.4

[gkronber]

r5551: removed ternary operator in LimitToRange extension as requested by mkommend. Removed range limitation in evaluators calculating the squared Pearson's r.

[mkommend]

r5552: Refactored Dataset.

[mkommend]

r5554: Refactored RegressionProblemData and related classes.

[gkronber]

r5553: worked on analyzers

[gkronber]

r5555: implemented variable frequency analyzer for symbolic data analysis

[gkronber]

r5556: renamed analyzers

[gkronber]

r5557: implemented single-objective and multi-objective training best solution analyzers for symbolic data analysis plugins.

[mkommend]

r5559: worked on different ProblemData classes.

[gkronber]

r5564: fixed bugs in online evaluators introduced with r5559

[gkronber]

r5565: fixed minor bug in data analysis problem data.

[gkronber]

r5567: improved ChangeNodeTypeManipulation operator to make unit test succeed.

[gkronber]

r5569: improved implementation of ReplaceBranchManipulation operator to make unit test succeed

[gkronber]

r5570: fixed bug in dataset

[gkronber]

r5571: ported interpreter and simplifier for symbolic data analysis expression trees to version 3.4.0 and fixed interpretation of 'not' symbol.

[gkronber]

r5572:5573: pulled external evaluation plugin to version 3.4

[gkronber]

r5574:

• added test projects for data-analysis and symbolic data-analysis plugin.
• copied grammars to version 3.4.
• fixed bugs in interpretation and simplification of 'not' symbols.

[gkronber]

r5575: set svn:ignore properties

[mkommend]

r5577: worked on data analysis problem representation.

[mkommend]

r5578: Added IHeuristicOptimizationProblem and adapted all according classes.

[mkommend]

r5579: Adapted new ProblemData classes.

[mkommend]

r5580: Added single and multi-objective SymbolicDataAnalysisProblem classes and adapted generic type parameter.

[mkommend]

r5582: Added IFixedValueParameter.

[mkommend]

r5584: Corrected minor flaw in CheckedItemCollection and CheckedItemList.

[mkommend]

r5585: Corrected FixedValueParameter.

[mkommend]

r5586: Adapated DataAnalysisProblemData as well as RegressionProblemData to use parameters.

[mkommend]

r5590: removed not used files.

[mkommend]

r5600: Added new plugin DataAnalysis.Views-3.4.

[mkommend]

r5601: Finished work on ProblemData.

[gkronber]

r5607: worked on data analysis solutions and validation best analyzers (solution does not build).

[gkronber]

r5613: fixed compilation errors after adding implementations for symbolic data analysis solutions and validation analyzers.

[gkronber]

r5615: changed order of specialization of algorithms to Algorithm -> EngineAlgorithm -> HeuristicOptimizationAlgorithm.

[gkronber]

r5616: fixed parameter name bug in SymbolicDataAnalysisEvaluator.

[gkronber]

r5617: added implementation of linear regression as FixedDataAnalysisAlgorithm

[mkommend]

r5618: Added symbolic data analysis problems.

[gkronber]

r5620: implemented base classes for data analysis, classification, and regression solutions.

[mkommend]

r5623: Added possibility to import data from csv files.

[gkronber]

r5624: renamed interface for interpreter, worked on solutions and models and implemented SVM regression.

[mkommend]

r5625: Reorganized branch and removed CreateAble-Attribute from outdated classes.

[gkronber]

r5626: implemented support vector classification algorithm.

[mkommend]

r5630: Added bounded mean squared error evaluator.

[mkommend]

r5635: Removed ReadOnlySymbol and changed ADF-Symbols.

[gkronber]

r5644: introduced HeuristicOptimizationEngineAlgorithm

[gkronber]

r5649:

• implemented classes for classification based on a discriminant function and thresholds
• implemented interfaces and base classes for clustering.

[gkronber]

r5650: implemented recursive collection of result collections into runs in algorithm.

[gkronber]

r5651: implemented wrapper classes for k-Means clustering in alglib.

[gkronber]

r5657: Implemented calculation of thresholds for discriminant function classification solutions.

[gkronber]

r5658: implemented wrapper for LDA (linear discriminant analysis) implemented in alglib.

[gkronber]

r5659: fixed bugs in CrossValidation

[gkronber]

r5661: changed FixedDataAnalysisAlgorithm to start asynchronously (taken from Engine)

[gkronber]

r5662: added ensemble models for regression and classification.

[gkronber]

r5663:

• ported estimated values view for classification and regression solutions
• ported scatter plot and line chart for regression solutions.

[gkronber]

r5664: ported ROC, confusion matrix and discriminant function classification views and fixed bug in threshold calculation.

[gkronber]

r5678: Worked on calculation of thresholds for classification solutions based on discriminant functions.

[gkronber]

r5680: minor change in calculation of normal distribution PDF cutting points.

[gkronber]

r5681: refactored threshold calculators.

[gkronber]

r5685:

• implemented validation best solution analyzers for symbolic classification and regression
• added analyzers to symbolic data analysis problem classes
• changed details of parameter wiring in problem classes.

[mkommend]

r5686: Finally added results from the grammar refactoring.

[mkommend]

r5687: Moved outdated plugins to separated solution folder.

[mkommend]

r5688: Added names and descriptions to grammars and adapted their view.

[gkronber]

r5689: fixed persistence problems.

[gkronber]

r5690: removed SupportVectorIndizes field from libSVM models.

[mkommend]

r5691: Fixed error in grammars.

[gkronber]

r5693: added plugin for symbolic data analysis views.

[gkronber]

r5694: ported views for support vector machine solutions and models.

[mkommend]

r5695: Implemented persistence of grammars. Currently it does not work, because the Tuple type is not serializable (see #1442).

[gkronber]

r5696: fixed bug in SVM solution view

[mkommend]

r5697: Added new assemblies to HL3.3 tests project and fixed minor errors regarding the visibility of the cloning and storable ctor.

[gkronber]

r5699: added interactive simplifier view for symbolic regression models.

[mkommend]

r5712:Fixed persistence of SymbolicExpressionGrammars.

[mkommend]

r5713: Small change in TupleSerializer.

[mkommend]

r5715: Removed unused files.

[mkommend]

r5716: Updated estimation limits on symbolic data analysis problem creation.

[gkronber]

r5717: Implemented interactive simplifier views for symbolic classification and regression.

[mkommend]

r5719: Refactored MultiManipulators of SymbolicExpressionTreeEncoding.

[gkronber]

r5720: added upper and lower estimation bounds for symbolic classification and regression models.

[gkronber]

r5722:

• fixed evaluator call from validation analyzers
• fixed bugs in interactive simplifier view
• added apply linear scaling flag to analyzers.

[mkommend]

r5726: Fixed grammars to allow ADF's.

[mkommend]

r5727: Adapted ProbabilisticTreeCreator.

[mkommend]

r5728: Corrected deserialization of SymbolicExpressionTreeNode.

[gkronber]

r5729:

• moved linear scaling method into symbolic regression model
• fixed bug in interactive solution simplifier

[mkommend]

r5730: Reorganized Problems.DataAnalysis.

[mkommend]

r5731: Commited remaining changes from project reorganization.

[gkronber]

r5732: fixed bug in variable frequency analyzer.

[mkommend]

r5733: Corrected problem interfaces & unified naming of subtrees.

[gkronber]

r5734: changed folder name to 'Outdated Plugins'

[gkronber]

r5736: implemented linear scaling for classification solutions, fixed bugs interactive simplifier view for classification solutions.

[gkronber]

r5737: removed unused classes

[gkronber]

r5742: fixed grammar cloning bug, improved analyzers for multi objective symbolic data analysis problems.

[gkronber]

r5745: ported formatters for data analysis expressions.

[gkronber]

r5746: bug fixes and minor changes to improve performance in online evaluators.

[gkronber]

r5747:

• corrected naming of bounded evaluator
• added base classes for single objective and multi objective validation analzers
• added overfitting analyzers for symbolic regression and classification.

[gkronber]

r5748: added calculation of variable impacts to variable frequency analyzer.

[gkronber]

r5749: minor changes in SymbolicDataAnalysisExpressionTreeInterpreter to prepare for integration of interval arithmetic.

[gkronber]

r5750:

• implemented variable view that allows to add and remove variable strings (ported from external evaluation views)
• implemented formatter and grammar for external evaluation plugin
• fixed bug in symbolic expression view
• fixed bug in run collection variable impact view
• removed unused plugin HeuristicLab.Problems.ExternalEvaluation.GP.Views
• renamed files.

[mkommend]

r5757: Implemented StringConvertibleValueTuple, IntRange, DoubleRange and the according views.

[mkommend]

r5758: Removed unused project Problems.ExternalEvaluation.GP.Views-3.4 from the solution file.

[mkommend]

r5759:

• Worked on IntRange and DoubleRange
• Updated evaluators, analyzers, problems and problem data to use IntRanges
• Removed properties to access the value of LookupParameter
• Corrected files.txt

[mkommend]

r5770: Changed data type of estimation limits and corrected some bugs.

[mkommend]

r5771: Updated overfitting analyzers to do nothing if validation partition is not set.

[mkommend]

r5772: Fixed bug in CrossValidation.

[mkommend]

r5774: Corrected ItemImage of StringConvertibleValueTypeTuple.

[mkommend]

r5775: Renamed ModelSelectionPartition to ValidationPartition.

[mkommend]

r5776: Removed unused class SymbolicDataAnalysisSolutionCreator.

[mkommend]

r5777: Corrected plugin dependencies.

[gkronber]

r5788: fixed bug in lagged variable view.

[mkommend]

r5791: Corrected deserialization of StringConvertibleValueTuple.

[mkommend]

r5792: Corrected ADFs and adapted unit tests.

[mkommend]

r5796: Merged trunk changes into branch.

[mkommend]

r5802: Added NonDiscoverableType attribute to outdated algorithms and problems.

[mkommend]

r5803: Corrected spacing of problem names.

[mkommend]

r5804: Removed RunCollectionVariableImpactView from outdated plugin, because it is present in the new plugin and would be shown twice otherwise.

[mkommend]

r5805: Removed RunCollectionVariableImpactView from outdated plugin, because it is present in the new plugin and would be shown twice otherwise.

[mkommend]

r5806: Updated samples partition for cross validation if a new problem data is imported.

[mkommend]

r5807: Enabled moving of thresholds in DiscrimantFunctionClassifcationSolutionView.

[mkommend]

r5808: Updated directories and removed old files.

[mkommend]

r5809: Reintegrated branch into trunk.

[mkommend]

r5810: Deleted DataAnalysis Refactoring branch.

[mkommend]

r5812: Updated ProbabilisticTreeCreatorTest.

[mkommend]

r5813: Fixed drag and drop on CrossValidationView.

[gkronber]

r5815: created branch for solution ensembles.

[gkronber]

r5817: fixed bug in LaTeX formatter.

[gkronber]

r5818: fixed a problem with scaling of regression and classification solutions (moved scale method out of solution into the model because of leaky abstraction).

[gkronber]

r5822: added unit test for symbolic regression with standard GP.

[gkronber]

r5823: fixed minor issues in symbolic data analysis classes.

[mkommend]

r5825: Added new projects to HL-3.3.Tests.

[mkommend]

r5835: Changed default formatter of SymbolicExpressionView and hid the maximization parameter for single objective symbolic data analysis problems.

[mkommend]

r5836: Fixed Symbolic Regression Unit Test.

[gkronber]

r5841: added missing wiring of SymbolicExpressionTreeGrammarParameter in SymbolicDataAnalysisProblem

[mkommend]

r5847: Adapted data analysis classes to new parameter ctors.

[gkronber]

r5849: fixed bug in threshold calculation and added missing wiring of model changed event in DiscriminantFunctionClassificationSolution.

[mkommend]

r5854: Hid additional parameters of symbolic datanalysis problem.

[mkommend]

r5857: Corrected SymbolicRegressionUnitTest and fixed typo in validation parameter description of SymbolicDataAnalysisProblem.

[mkommend]

r5860: Adapted PluginDescriptions and AssemblyDescriptions for new plugins.

[mkommend]

r5861: Corrected cloning bug in SupportVectorMachineModel.

[mkommend]

r5862: Removed maximum value for number of workers in CrossValidationView as this could cause problems when loading files executed on a different machine.

[mkommend]

r5863: Added NonDiscoverableType attribute to outdated analyzers.

[gkronber]

r5869: minor improvements in plugin descriptions.

[mkommend]

r5879: Deleted unnecessary files.

[gkronber]

r5882: renamed parameter and updated all validation analyzers to leave out test samples if the validation partition overlaps with the test partition.

[mkommend]

r5883: Adapted interfaces of data analysis classes.

[gkronber]

r5884: updated symbolic regression sample for unit test.

[gkronber]

r5885: added R² and MSE as results for discriminant function classification solutions.

[mkommend]

r5886: Corrected bugs in CrossValidation when using evolutionary algorithms.

[mkommend]

r5887: Reset ValidationPartition when new problem data was loaded.

[gkronber]

r5888: sorted deploymentitems in builder test settings alphabetically and added new plugins added in data analysis refactoring.

[gkronber]

r5889: deleted unused file SymbolicDataAnalysisSolution.cs and fixed compiler warning.

[gkronber]

r5890: fixed assembly name and namespace for symbolic data analysis tests assembly.

[gkronber]

r5895: adapted limit for failed events in unit test for ArgumentCreater

[mkommend]

r5907: Removed calls to IEnumerable.Count() in performance critical sections.

[gkronber]

r5909: made data analysis solutions storable.

[mkommend]

r5914: Changed DataAnalysisSolutions and -Models and updated GenerateRowsToEvaluate method in SymbolicDataAnalysisEvaluator.

[gkronber]

r5915: fixed bug in symbolic data analysis expression tree simplifier and added test cases.

[gkronber]

r5916: changed crossover operator for symbolic expression trees to allow extensions and reductions of the number of subtrees for variable-arity symbols.

[gkronber]

r5917: minor bug fix in all architecture altering operators test.

[gkronber]

r5918: minor changes in all architecture altering operators unit test.

[mkommend]

r5921: Fixed broken SymbolicRegressionUnitTest.

[gkronber]

r5922: fixed bug in variable frequency analyzer (when lagged variables are not aggregated)

[gkronber]

r5923: fixed bugs in symbolic data analysis expression tree interpreter regarding evaluation of models with lagged variables and dynamic symbols.

[gkronber]

r5924: fixed another bug in variable frequency analyzer.

[gkronber]

r5925: fixed bugs in symbolic expression tree operators.

[gkronber]

r5926: changed views for time lag and lagged variable symbols to allow a lag value of zero

[mkommend]

r5936: Removed asserts from SymbolicDataAnalysisExpressionTreeSimplifier.

[mkommend]

r5937: Fixed broken SymbolicRegressionUnitTest again.

[mkommend]

r5944: Added unit tests for the Problems.DataAnalysis-3.4 project.

[mkommend]

r5962: Fixed OnlineNormalizedMeanSquaredErrorCalculator and added results description in RegressionSolution.

[gkronber]

r5971: added a parameter to the symbol frequency analyzer in symbolic expression tree encoding which indicates if the frequencies of symbols with the same name but different number of subtrees should be aggregated.

[mkommend]

r5983: Added after deserialization hook to SymbolicExpressionSymbolFrequencyAnalyzer for newly added parameter.

[gkronber]

r5987: improved memory usage of symbolic data analysis expression tree interpreter.

[gkronber]

r5993: fixed issues in interactive simplifier views for regression and classification solutions regarding automatic folding of nodes without impact on the solution quality and problems with ADFs.

[gkronber]

Further related changes should be tracked in more specific tickets.