Implement views for decision tree models (random forest and gradient boosted trees)

[gkronber]

It would be nice if we could inspect decision tree models

[gkronber]

r14345: implemented methods to generate symbolic expression tree solutions for decision tree models (random forest and gradient boosted) as well as views which make it possible to inspect each of the individual trees in a GBT and RF solution

[gkronber]

r14346: removed unnecessary plugin dependency

[mkommend]

Bug: At least RF models are not stored directly, but rather recreated when necessary. However the new model view does not trigger this recreation and throws an out-of-bounds exception if a deserialized model is displayed.

[gkronber]

Replying to mkommend:

Bug: At least RF models are not stored directly, but rather recreated when necessary. However the new model view does not trigger this recreation and throws an out-of-bounds exception if a deserialized model is displayed.

Thanks. Should be fixed with r14368.

[mkommend]

Review comments:

• Double click in the list view does not open the symbolic regression solution. (added with r15103)
• The model view is the last entry in the solution view.
• Why are evaluation views misused to display the individual trees. Would not it be possible to create a view for IRandomForestModel / GBTModel? (gkronber: model views have been introduced with r15105:15106 and r15123:15124 additionally to model evaluation views)
• Why are GBTSolutionView and RF Regression/Classifation View necessary? Only to remove the menu entry? (gkronber: solution views have been removed with r15105:15106 and r15123:15124)
• Excel format for the symbolic models does not work correctly. There is no infinity symbol in excel and additionally 1/(1+EXP(-Infinity * ($D1-44)) is not easy to interpret correctly. Especially for RF where ignoreSlope == true a standard if-condition should be exported. (fixed with r15104)
• Depending on whether ignoreSlope is set the condition is negated (> changes to <).
• Slope of +/- infinity has the same meaning as ignoreSlope=true and not slope == 0 (VariableConditionTreeNode.ToString).
• VariableConditionView should show the ignoreSlope flag
• IMHO IgnoreSlope should be the default for variable condition symbols.

gkronber: After discussion we decided it would be best to introduce a new type of symbol especially for this (split var <= threshold). This also means that we need to extend interpreters and formatters to support this symbol.

[mkommend]

r14839: Adapated content attribute of solution views for GBT and RF.

This change has been necessary to show the solution name text box from the NamedDataAnalysisSolutionView.

[gkronber]

r15103: added event-handler for double click events to open the clicked model in a new view

[gkronber]

r15104: changed Excel-formatter for VariableCondition symbols to produce a correct output also for GBT and RF models

[gkronber]

Replying to mkommend:

Review comments:

• The model view is the last entry in the solution view.
• Why are evaluation views misused to display the individual trees. Would not it be possible to create a view for IRandomForestModel / GBTModel?

This is not possible as the views for the individual trees need the problem data to calculate the results for the indiviual trees. However it would be possible to provide both views. Whereas, the model view only shows the trees and the other view is renamed.

• Why are GBTSolutionView and RF Regression/Classifation View necessary? Only to remove the menu entry?

Yes, but probably not anymore if we keep both entries.

[gkronber]

r15105:15106: added a model view for GBT models and renamed the existing evaluation view accordingly. Deleted GBT solution view because it is not necessary anymore.

[gkronber]

r15123 renamed RandomForestModelView -> RandomForestModelEvaluationView

[gkronber]

r15124: removed solution views for RFRegression and RFClassification. Added model view for IRandomForestModels

[gkronber]

r15127: merged r14345, r14346, r14368 from trunk to stable (only partially merged to stable!)

Remaining changesets should be merged after r14826 in #2650.

[gkronber]

r15134: merged r14839,r15103:15106,r15123,r15124 from trunk to stable (all changesets merged)