Package weka.experiment
package weka.experiment
-
ClassDescriptionTakes the results from a ResultProducer and submits the average to the result listener.A SplitEvaluator that produces results for a classification scheme on a nominal class attribute.Interface to something that can accept remote connections and execute a task.SplitEvaluator that produces results for a classification scheme on a nominal class attribute, including weighted misclassification costs.Generates for each run, carries out an n-fold cross-validation, using the set SplitEvaluator to generate some results.Carries out one split of a repeated k-fold cross-validation, using the set SplitEvaluator to generate some results.Takes results from a result producer and assembles them into comma separated value form.Takes results from a result producer and sends them to a database.Examines a database and extracts out the results produced by the specified ResultProducer and submits them to the specified ResultListener.DatabaseUtils provides utility functions for accessing the experiment database.A SplitEvaluator that produces results for a density based clusterer.Holds all the necessary configuration information for a standard type experiment.Loads the external test set and calls the appropriate SplitEvaluator to generate some results.
The filename of the test set is constructed as follows:
<dir> + / + <prefix> + <relation-name> + <suffix>
The relation-name can be modified by using the regular expression to replace the matching sub-string with a specified replacement string.Convert the results of a database query into instances.An interface implemented by InstanceQuery and any user class that is to be passed as the first argument to InstanceQuery.retrieveInstances(InstanceQueryAdapter, ResultSet).Outputs the received results in arff format to a Writer.Tells a sub-ResultProducer to reproduce the current run for varying sized subsamples of the dataset.OutputZipper writes output to either gzipped files or to a multi entry zip file.Behaves the same as PairedTTester, only it uses the corrected resampled t-test statistic.A class for storing stats on a paired comparison (t-test and correlation)A class for storing stats on a paired comparison.Calculates T-Test statistics on data stored in a set of instances.Stores information on a property of an object: the class of the object with the property; the property descriptor, and the current value.Generates a single train/test split and calls the appropriate SplitEvaluator to generate some results.A SplitEvaluator that produces results for a classification scheme on a numeric class attribute.A general purpose server for executing Task objects sent via RMI.Holds all the necessary configuration information for a distributed experiment.Class encapsulating information on progress of a remote experimentInterface for classes that want to listen for updates on RemoteExperiment progressClass to encapsulate an experiment as a task that can be executed on a remote host.Interface for objects able to listen for results obtained by a ResultProducerThis matrix is a container for the datasets and classifier setups and their statistics.Generates the matrix in CSV ('comma-separated values') format.Generates output for a data and script file for GnuPlot.Generates the matrix output as HTML.Generates the matrix output in LaTeX-syntax.Generates the output as plain text (for fixed width fonts).Only outputs the significance indicators.This interface defines the methods required for an object that produces results for different randomizations of a dataset.Interface to objects able to generate a fixed set of results for a particular split of a dataset.A class to store simple statistics.Interface to something that can be remotely executed as a task.A class holding information for tasks being executed on RemoteEngines.Interface for different kinds of Testers in the Experimenter.