File
File#
weka/gui/explorer/Explorer.props
Description#
This props file determines what schemes and options are initially set in the Explorer.
Version#
-
3.5.3
Fields#
Preprocess panel#
- InitGenericObjectEditorFilter
if set to true the Capabilities filters in the GOE will be initialized based on the full dataset that has been loaded into the Explorer otherwise only the header
- Tabs
Lists all the tabs that should be displayed in the Explorer. Apart from the Preprocess panel itself, all other panels are basically plugins. See the Adding tabs in the Explorer article for more details on adding custom panels.
-
InitialDirectory (> 3.6.0, developer version)
Defines the initial directory for opening datasets in the Preprocess panel. The following placeholders are recognized (work across platforms):
- %t - the temp directory
- %h - the user's home directory
- %c - the current directory (the default setting)
- %% - gets replaced by a single percentage sign
-
enableUndo (> 3.6.5, > 3.7.4)
Enable/disable the creation of undo files (default is enabled)
-
undoDirectory (> 3.6.5, > 3.7.4)
Specify the directory to use for saving undo files The following placeholders are recognized (work across platforms):
- %t - the temp directory
-
Filter
the filter to use, none if left empty
Classify panel#
- Classifier
the classifier to use
- ClassifierTestMode
the default test mode in the classify tab
- 1 - cross-validation (default)
- 2 - percentage split
- 3 - use training set
- 4 - supplied test set
- ClassifierCrossvalidationFolds
the default number of folds for CV
- ClassifierCostSensitiveEval
whether the evaluation of the classifier is done cost-sensitively a cost matrix still has to be provided!
- ClassifierOutputConfusionMatrix
whether the confusion matrix is output
- ClassifierOutputEntropyEvalMeasures
whether the entropy based evaluation measures of the classifier model are output
- ClassifierOutputModel
whether the classifier model is output
- ClassifierOutputPerClassStats
whether additional per-class stats of the classifier model are output
- ClassifierOutputPredictions
whether the predictions of the classifier output as well
- ClassifierPercentageSplit
the default percentage split in %
- ClassifierPreserveOrder
whether the order is preserved in case of percentage split
- ClassifierRandomSeed
the default random seed
- ClassifierStorePredictionsForVis
whether the predictions of the classifier are stored for visulization purposes
- ClassifierOutputSourceCode (> 3.5.5)
whether to output Java source code for classifiers that implement the weka.classifiers.Sourcable interface
- ClassifierSourceCodeClass (> 3.5.5)
the default classname of the generated Java source code
- ClassifierErrorsPlotInstances (> 3.7.0)
the default classname for the class generating the plot instances of the classifier errors
- ClassifierErrorsMinimumPlotSizeNumeric (> 3.7.0)
the minimum size for the crosses that display the classifier errors for numeric class attributes
- ClassifierErrorsMaximumPlotSizeNumeric (> 3.7.0)
the maximum size for the crosses that display the classifier errors for numeric class attributes
Cluster panel#
- Clusterer
the clusterer to use
- ClustererTestMode
the default test mode
- 2 - percentage split
- 3 - use training set (default)
- 4 - supplied test set
- 5 - classes to clusters evaluation
- ClustererStoreClustersForVis
whether the clusters are stored for visualization purposes
- Associations panel
- Associator
the default associator
- Attribute selection panel
- ASEvaluation
the default attribute evaluator
- ASSearch
the default attribute selection search scheme
- ASTestMode
the default test mode
- 0 - use full training set (default)
- 1 - cross-validation
- ASCrossvalidationFolds
the default number of folds for CV
- ASRandomSeed
the default random seed