weka.core

Interface BatchPredictor

All Known Implementing Classes:

AbstractClassifier, AdaBoostM1, AdditiveRegression, AttributeSelectedClassifier, Bagging, BayesNet, BayesNetGenerator, BIFReader, ClassificationViaRegression, CostSensitiveClassifier, CVParameterSelection, DecisionStump, DecisionTable, EditableBayesNet, FilteredClassifier, GaussianProcesses, GeneralRegression, HoeffdingTree, IBk, InputMappedClassifier, IteratedSingleClassifierEnhancer, IterativeClassifierOptimizer, J48, JRip, KStar, LinearRegression, LMT, LMTNode, Logistic, LogisticBase, LogitBoost, LWL, M5Base, M5P, M5Rules, MultiClassClassifier, MultiClassClassifierUpdateable, MultilayerPerceptron, MultipleClassifiersCombiner, MultiScheme, NaiveBayes, NaiveBayesMultinomial, NaiveBayesMultinomialText, NaiveBayesMultinomialUpdateable, NaiveBayesUpdateable, NeuralNetwork, OneR, ParallelIteratedSingleClassifierEnhancer, ParallelMultipleClassifiersCombiner, PART, PMMLClassifier, PreConstructedLinearModel, RandomCommittee, RandomForest, RandomizableClassifier, RandomizableFilteredClassifier, RandomizableIteratedSingleClassifierEnhancer, RandomizableMultipleClassifiersCombiner, RandomizableParallelIteratedSingleClassifierEnhancer, RandomizableParallelMultipleClassifiersCombiner, RandomizableSingleClassifierEnhancer, RandomSubSpace, RandomTree, Regression, RegressionByDiscretization, REPTree, RuleNode, RuleSetModel, SerializedClassifier, SGD, SGDText, SimpleLinearRegression, SimpleLogistic, SingleClassifierEnhancer, SMO, SMOreg, Stacking, SupportVectorMachineModel, TreeModel, Vote, VotedPerceptron, WeightedInstancesHandlerWrapper, ZeroR

public interface BatchPredictor

Interface to something that can produce predictions in a batch manner when presented with a set of Instances.

Version:

$Revision: 11958 $

Author:

Mark Hall (mhall{[at]}pentaho{[dot]}com)

Method Summary

Modifier and Type	Method and Description
double[][]	distributionsForInstances(Instances insts) Batch scoring method
String	getBatchSize() Get the batch size to use.
boolean	implementsMoreEfficientBatchPrediction() Returns true if this BatchPredictor can generate batch predictions in an efficient manner.
void	setBatchSize(String size) Set the batch size to use.

Method Detail

setBatchSize

void setBatchSize(String size)

Set the batch size to use. The implementer will prefer (but not necessarily expect) this many instances to be passed in to distributionsForInstances().

Parameters:

size - the batch size to use

getBatchSize

String getBatchSize()

Get the batch size to use. The implementer will prefer (but not necessarily expect) this many instances to be passed in to distributionsForInstances(). Allows the preferred batch size to be encapsulated with the client.

Returns:

the batch size to use

distributionsForInstances

double[][] distributionsForInstances(Instances insts)
                              throws Exception

Batch scoring method

Parameters:

insts - the instances to get predictions for

Returns:

an array of probability distributions, one for each instance

Throws:

Exception - if a problem occurs

implementsMoreEfficientBatchPrediction

boolean implementsMoreEfficientBatchPrediction()

Returns true if this BatchPredictor can generate batch predictions in an efficient manner.

Returns:

true if batch predictions can be generated efficiently