AbstractCallableWekaClassifierEvaluator |
Ancestor for classifier evaluators that make use of a callable classifier.
|
AbstractCallableWekaClustererEvaluator |
Ancestor for clusterer evaluators that make use of a callable clusterer.
|
AbstractInstanceGenerator<T extends adams.data.container.DataContainer> |
Ancestor for transformers that turn data containers into WEKA Instance
objects.
|
AbstractProcessWekaInstanceWithModel<T> |
Ancestor for transformers that user models for processing Instance objects,
e.g., classifiers making predictions.
|
AbstractWekaClassifierEvaluator |
Ancestor for transformers that evaluate classifiers.
|
AbstractWekaModelReader |
Ancestor for actors that deserialize models.
|
AbstractWekaPredictionsTransformer |
Ancestor for transformers that convert the predictions stored in an
Evaluation object into a different format.
|
WekaAccumulatedError |
Generates plot containers from an evaluation object's predictions.
|
WekaAccumulatedError.SortablePrediction |
Container for a classifier prediction, used for sorting.
|
WekaAggregateEvaluations |
Aggregates incoming weka.classifiers.Evaluation objects and forwards the current aggregated state.
|
WekaAttributeIterator |
Iterates through all attributes of a dataset and outputs the names.
The attributes can be limited with the range parameter and furthermore with the regular expression applied to the names.
Instead of outputting the names, it is also possible to output the 1-based indices.
|
WekaAttributeSelection |
Performs attribute selection on the incoming data.
In case of input in form of a class adams.flow.container.WekaTrainTestSetContainer object, the train and test sets stored in the container are being used.
NB: In case of cross-validation no reduced or transformed data can get generated!
Input/output:
- accepts:
weka.core.Instances
adams.flow.container.WekaTrainTestSetContainer
- generates:
adams.flow.container.WekaAttributeSelectionContainer
Container information:
- adams.flow.container.WekaTrainTestSetContainer: Train, Test, Seed, FoldNumber, FoldCount, Train original indices, Test original indices
- adams.flow.container.WekaAttributeSelectionContainer: Train, Reduced, Transformed, Test, Test reduced, Test transformed, Evaluation, Statistics, Selected attributes, Seed, FoldCount
|
WekaAttributeSelectionSummary |
Outputs a summary string of the attribute selection.
|
WekaBootstrapping |
Performs bootstrapping on the incoming evaluation and outputs a spreadsheet where each row represents the results from bootstrapping sub-sample.
|
WekaChooseAttributes |
Lets the user select attributes interactively to use down the track.
Internally, a weka.filters.unsupervised.attribute.Remove WEKA filter is constructed from the selection, to remove the attributes that the user didn't select.
|
WekaClassifierInfo |
Outputs information of a trained weka.classifiers.Classifier object.
|
WekaClassifierOptimizer |
Evaluates a classifier optimizer on an incoming dataset.
|
WekaClassifierRanker |
Performs a quick evaluation using cross-validation on a single dataset (or evaluation on a separate test set if the number of folds is less than 2) to rank the classifiers received on the input and forwarding the x best ones.
|
WekaClassifierRanker.RankingJob |
A job class specific to ranking classifiers.
|
WekaClassifierSetupProcessor |
Applies the specified processor to the incoming array of classifiers, e.g., for generating new or filtered setups.
|
WekaClassifying |
Uses a serialized model to perform predictions on the data being passed through.
The following order is used to obtain the model (when using AUTO):
1.
|
WekaClassSelector |
Sets the class index.
|
WekaClusterAssignments |
Outputs the cluster assignments from the evaluation.
|
WekaClustererInfo |
Outputs information of a trained weka.clusterers.Clusterer object.
|
WekaClustererPostProcessor |
Applies the specified post-processor to the cluster container (adams.flow.container.WekaModelContainer)
See also:
adams.flow.transformer.WekaTrainClusterer
Input/output:
- accepts:
adams.flow.container.WekaModelContainer
- generates:
adams.flow.container.WekaModelContainer
Container information:
- adams.flow.container.WekaModelContainer: Model, Header, Dataset
|
WekaClusterEvaluationSummary |
Generates a summary string of the weka.clusterers.ClusterEvaluation objects that it receives.
|
WekaClustering |
Uses a serialized model to cluster data being passed through.
The following order is used to obtain the model (when using AUTO):
1.
|
WekaCrossValidationClustererEvaluator |
Cross-validates a clusterer on an incoming dataset.
|
WekaCrossValidationEvaluator |
Cross-validates a classifier on an incoming dataset.
|
WekaCrossValidationSplit |
Generates train/test pairs like during a cross-validation run.
|
WekaDatasetsMerge |
Merges 2 or more datasets into a single dataset, under a selectable merge method.
|
WekaDatasetSplit |
Splits the incoming dataset into sub-sets using the specified splitter.
|
WekaEnsembleGenerator |
Uses the specified generator to create ensembles from the incoming data.
|
WekaEvaluationInfo |
Outputs information about a Weka weka.classifiers.Evaluation object.
|
WekaEvaluationPostProcessor |
Applies the specified post-processor to the incoming Evaluation data.
|
WekaEvaluationSummary |
Generates a summary string of the weka.classifiers.Evaluation objects that it receives.
|
WekaEvaluationValuePicker |
Picks a specific value from an evaluation object.
|
WekaEvaluationValues |
Generates a spreadsheet from statistics of an Evaluation object.
|
WekaExperiment |
Represents a Weka experiment, stored in a file.
|
WekaExperimentEvaluation |
Generates evaluation output of an experiment that was run previously.
|
WekaExperimentExecution |
Executes an experiment.
|
WekaExperimentFileReader |
Loads an experiment file.
|
WekaExtractArray |
Extracts a column or row of data from a weka.core.Instances or SpreadSheet object.
Only numeric columns can be returned.
|
WekaExtractPLSMatrix |
Transformer that allows the extraction of internal PLS filter/classifier matrices, forwarding them as spreadsheets.
|
WekaFileReader |
Reads any file format that Weka's converters can handle and returns the full dataset or single weka.core.Instance objects.
|
WekaFilter |
Filters Instances/Instance objects using the specified filter.
When re-using a trained filter, ensure that 'initializeOnce' is checked.
The following order is used to obtain the model (when using AUTO):
1.
|
WekaFilter.BatchFilterJob |
|
WekaGenericPLSMatrixAccess |
Transformer that allows the extraction of internal PLS filter/classifier matrices, forwarding them as spreadsheets.
See the respective PLS implementation for details on available matrix names (derived from: weka.filters.supervised.attribute.pls.AbstractPLS)
Input/output:
- accepts:
weka.classifiers.Classifier
weka.filters.Filter
weka.core.GenericPLSMatrixAccess
adams.flow.container.WekaModelContainer
- generates:
adams.data.spreadsheet.SpreadSheet
Container information:
- adams.flow.container.WekaModelContainer: Model, Header, Dataset
|
WekaGeneticAlgorithm |
Applies the genetic algorithm to the incoming dataset.
Forwards the best setup(s) after the algorithm finishes.
A callable sink can be specified for receiving intermediate performance results.
|
WekaGeneticAlgorithmInitializer |
Populates a adams.flow.container.WekaGeneticAlgorithmInitializationContainer container from the data obtained from the incoming setup (in properties format, can be gzip compressed).
|
WekaGetCapabilities |
Retrieves the capabilities of a weka.core.CapabilitiesHandler (eg filter or classifier) and forwards them.
|
WekaGetInstancesValue |
Retrieves a value from a WEKA Instances object.
Notes:
- date and relational values are forwarded as strings
- missing values are output as '?' (without the single quotes)
Input/output:
- accepts:
weka.core.Instances
- generates:
java.lang.Double
java.lang.String
|
WekaGetInstanceValue |
Retrieves a value from a WEKA Instance object.
Notes:
- date and relational values are forwarded as strings
- missing values are output as '?' (without the single quotes)
- the 'attribute name' option overrides the 'index' option
Input/output:
- accepts:
weka.core.Instance
- generates:
java.lang.Double
java.lang.String
Valid options are:
|
WekaInstanceBuffer |
Can act in two different ways:
1.
|
WekaInstanceDumper |
Dumps weka.core.Instance objects into an ARFF file.
|
WekaInstanceEvaluator |
Adds a new attribute to the data being passed through (normally 'evaluation') and sets the value to the evaluation value returned by the chosen evaluator scheme.
|
WekaInstanceFileReader |
Loads a WEKA dataset from disk with a specified reader and passes on the adams.core.instance.Instance objects.
|
WekaInstancesAppend |
Creates one large dataset by appending all one after the other.
|
WekaInstancesHistogramRanges |
Outputs the ranges generated by adams.data.statistics.ArrayHistogram using the incoming weka.core.Instances object.
The actor just uses the internal format (double array) and does not check whether the attributes are actually numeric.
|
WekaInstancesInfo |
Outputs statistics of a weka.core.Instances object.
FULL_ATTRIBUTE and FULL_CLASS output a spreadsheet with detailed attribute statistics.
|
WekaInstancesMerge |
Merges multiple datasets, either from file or using Instances/Instance objects.
If no 'ID' attribute is named, then all datasets must contain the same number of rows.
Attributes can be excluded from ending up in the final dataset via a regular expression.
|
WekaInstancesStatistic |
Generates statistics from a weka.core.Instances object.
The actor just uses the internal format (double array) and does not check whether the attributes are actually numeric.
|
WekaInstanceStreamPlotGenerator |
Generates plot containers from a range of attributes of the weka.core.Instance objects being passed through.
The generator merely uses the internal data representation for generating the Y value of the plot container.
|
WekaModelReader |
Actor for loading a model (classifier or clusterer).
|
WekaMultiLabelSplitter |
Splits a dataset containing multiple class attributes ('multi-label') into separate datasets with only a single class attribute.
|
WekaNearestNeighborSearch |
Outputs the specified number of nearest neighbors for the incoming Weka Instance.
The data used for the nearest neighbor search is either obtained from storage.
|
WekaNewInstance |
Creates a new weka.core.Instance-derived object, with all values marked as missing.
The class implementing the weka.core.Instance interface needs to have a constructor that takes the number of attributes as sole parameter.
|
WekaPackageManagerAction |
Applies the selected Weka Package Manager action to the incoming data and forwards the generated output.
|
WekaPredictionsToInstances |
Generates weka.core.Instances from the predictions of an Evaluation object.
|
WekaPredictionsToSpreadSheet |
Generates a SpreadSheet object from the predictions of an Evaluation object.
See also:
adams.flow.transformer.WekaSpreadSheetToPredictions
Input/output:
- accepts:
weka.classifiers.Evaluation
adams.flow.container.WekaEvaluationContainer
- generates:
adams.data.spreadsheet.SpreadSheet
Container information:
- adams.flow.container.WekaEvaluationContainer: Evaluation, Model, Prediction output, Original indices
|
WekaPrincipalComponents |
Performs principal components analysis on the incoming data and outputs the loadings and the transformed data as spreadsheet array.
Automatically filters out attributes that cannot be handled by PCA.
|
WekaRandomSplit |
Splits a dataset into a training and test set according to a specified split percentage.
|
WekaRegexToRange |
Produces a range string from a regular expression describing attributes.
|
WekaRelationName |
Deprecated. |
WekaRenameRelation |
Modifies relation names.
|
WekaReorderAttributesToReference |
Reorders the attributes of the Instance/Instances passing through according to the provided reference dataset (callable actor or reference file).
This ensures that the generated data always has the same structure as the reference dataset.
|
WekaRepeatedCrossValidationEvaluator |
Performs repeated cross-validation a classifier on an incoming dataset.
|
WekaRepeatedCrossValidationOutput |
Generates output from the incoming repeated cross-validation data.
|
WekaSetInstancesValue |
Sets a value in a WEKA Instances object.
Notes:
- relational values cannot be set
- '?' (without single quotes) is interpreted as missing value
Input/output:
- accepts:
weka.core.Instances
- generates:
weka.core.Instances
|
WekaSetInstanceValue |
Sets a value in a WEKA Instance.
Notes:
- relational values cannot be set
- '?' (without single quotes) is interpreted as missing value
Input/output:
- accepts:
weka.core.Instance
- generates:
weka.core.Instance
Valid options are:
|
WekaSplitGenerator |
WekaSpreadSheetToPredictions |
Turns the predictions stored in the incoming spreadsheet (actual and predicted) into a Weka weka.classifiers.Evaluation object.
For recreating the predictions of a nominal class, the class distributions must be present in the spreadsheet as well.
See also:
adams.flow.transformer.WekaPredictionsToSpreadSheet
Input/output:
- accepts:
adams.data.spreadsheet.SpreadSheet
- generates:
weka.classifiers.Evaluation
|
WekaStoreInstance |
Appends the incoming weka.core.Instance to the dataset in storage.
|
WekaStreamEvaluator |
Evaluates an incremental classifier on a data stream using prequential evaluation (first evaluate, then train).
|
WekaStreamFilter |
Filters Instance objects using the specified filter.
|
WekaSubsets |
Splits the dataset based on the unique values of the specified attribute: all rows with the same unique value form a subset.
|
WekaTestSetClustererEvaluator |
Evaluates a trained clusterer (obtained from input) on the dataset obtained from the callable actor.
If a class attribute is set, a classes-to-clusters evaluation is performed automatically
Input/output:
- accepts:
weka.clusterers.Clusterer
adams.flow.container.WekaModelContainer
- generates:
adams.flow.container.WekaClusterEvaluationContainer
Container information:
- adams.flow.container.WekaModelContainer: Model, Header, Dataset
- adams.flow.container.WekaClusterEvaluationContainer: Evaluation, Model, Log-likelohood
|
WekaTestSetEvaluator |
Evaluates a trained classifier (obtained from input) on the dataset obtained from the callable actor.
|
WekaTestSetEvaluator.EvaluateJob |
|
WekaTextDirectoryReader |
Loads all text files in a directory and uses the subdirectory names as class labels.
|
WekaTrainAssociator |
Trains an associator based on the incoming dataset and outputs the built associator alongside the training header and rules (in a model container)..
|
WekaTrainAssociator.TrainJob |
|
WekaTrainClassifier |
Trains a classifier based on the incoming dataset and outputs the built classifier alongside the training header (in a model container).
Incremental training is performed, if the input are weka.core.Instance objects and the classifier implements weka.classifiers.UpdateableClassifier.
|
WekaTrainClassifier.BatchTrainJob |
|
WekaTrainClusterer |
Trains a clusterer based on the incoming dataset and output the built clusterer alongside the training header (in a model container).
Incremental training is performed, if the input are weka.core.Instance objects and the clusterer implements weka.clusterers.UpdateableClusterer.
|
WekaTrainClusterer.BatchTrainJob |
|
WekaTrainTestSetClustererEvaluator |
Trains a clusterer on an incoming training dataset (from a container) and then evaluates it on the test set (also from a container).
The clusterer setup being used in the evaluation is a callable 'Clusterer' actor.
If a class attribute is set, a classes-to-clusters evaluation is performed automatically
Input/output:
- accepts:
adams.flow.container.WekaTrainTestSetContainer
- generates:
adams.flow.container.WekaClusterEvaluationContainer
Container information:
- adams.flow.container.WekaTrainTestSetContainer: Train, Test, Seed, FoldNumber, FoldCount
- adams.flow.container.WekaClusterEvaluationContainer: Evaluation, Model, Log-likelohood
|
WekaTrainTestSetEvaluator |
Trains a classifier on an incoming training dataset (from a container) and then evaluates it on the test set (also from a container).
The classifier setup being used in the evaluation is a callable 'Classifier' actor.
|
WekaTrainTestSetEvaluator.EvaluateJob |
|