ai.libs.jaicore.ml.tsc.classifier.neighbors

Class ShotgunEnsembleClassifier

java.lang.Object

ai.libs.jaicore.ml.tsc.classifier.ASimplifiedTSClassifier<java.lang.Integer>

ai.libs.jaicore.ml.tsc.classifier.neighbors.ShotgunEnsembleClassifier

public class ShotgunEnsembleClassifier
extends ASimplifiedTSClassifier<java.lang.Integer>

Implementation of Shotgun Ensemble Classifier as published in "Towards Time Series Classfication without Human Preprocessing" by Patrick Schäfer (2014). The Shotgun Classifier is based 1-NN and the Shotgun Distance. The Shotgun Ensemble Algorithm ShotgunEnsembleAlgoritm determines for specific window lengths the number of correct predicitions on the training data using the leave-one-out technique. The bestScore is the highest number of correct predicitions over all window lengths. Given a factor in (0,1], the window lengths where correct * factor > bestScore are used in an ensemble of Shotgun Classifiers to create an overall predicition.

Field Summary

Fields

Modifier and Type	Field and Description
protected int	bestScore The best score.
protected double	factor Factor used to determine whether or not to include a window length into the overall predicition.
protected NearestNeighborClassifier	nearestNeighborClassifier The nearest neighbor classifier used for prediction.
protected ShotgunDistance	shotgunDistance The Shotgun Distance used by the nearestNeighborClassifier.
protected int[]	targets Target values for the instances.
protected double[][]	values Value matrix containing the time series instances.
protected java.util.ArrayList<ai.libs.jaicore.basic.sets.Pair<java.lang.Integer,java.lang.Integer>>	windows Holds pairs of (number of correct predictions, window length) obtained in training phase.

Fields inherited from class ai.libs.jaicore.ml.tsc.classifier.ASimplifiedTSClassifier

classMapper, trained

Constructor Summary

Constructors

Constructor and Description
ShotgunEnsembleClassifier(int minWindowLength, int maxWindowLength, boolean meanNormalization, double factor) Creates a Shotgun Ensemble classifier.

Method Summary

Modifier and Type	Method and Description
protected java.util.Map<java.lang.Integer,java.lang.Integer>	calculateWindowLengthPredictions(double[] testInstance) Calculates predicitions for a test instance using 1NN with Shotgun Distance and different window lengths.
protected java.util.Map<java.lang.Integer,java.util.List<java.lang.Integer>>	calculateWindowLengthPredictions(TimeSeriesDataset dataset) Calculates predicitions for a test dataset using 1NN with Shotgun Distance and different window lengths.
ShotgunEnsembleLearnerAlgorithm	getLearningAlgorithm(TimeSeriesDataset dataset)
protected java.lang.Integer	mostFrequentLabelFromWindowLengthPredicitions(java.util.Map<java.lang.Integer,java.lang.Integer> windowLengthPredicitions) Returns the most frequent predicition given a Map of (window length, prediciton) pairs.
protected java.util.List<java.lang.Integer>	mostFrequentLabelsFromWindowLengthPredicitions(java.util.Map<java.lang.Integer,java.util.List<java.lang.Integer>> windowLengthPredicitions) Returns for each instance the most frequent predicitions as contained in a Map of (window length, list of prediciton for each instance) pairs.
java.lang.Integer	predict(double[] univInstance) Predicts on univariate instance.
java.lang.Integer	predict(java.util.List<double[]> multivInstance) Predicts on a multivariate instance.
java.util.List<java.lang.Integer>	predict(TimeSeriesDataset dataset) Predicts on a dataset.
protected void	setNearestNeighborClassifier(NearestNeighborClassifier nearestNeighborClassifier) Sets the nearest neighbor classifier, nearestNeighborClassifier.
protected void	setTargets(int[] targets) Sets the targets.
protected void	setValues(double[][] values) Sets the value matrix.
protected void	setWindows(java.util.ArrayList<ai.libs.jaicore.basic.sets.Pair<java.lang.Integer,java.lang.Integer>> windows) Sets the windows and also retreives and sets the @see #bestScore from these windows.

Methods inherited from class ai.libs.jaicore.ml.tsc.classifier.ASimplifiedTSClassifier

getClassMapper, isTrained, setClassMapper, train

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

factor

protected double factor

Factor used to determine whether or not to include a window length into the overall predicition.

values

protected double[][] values

Value matrix containing the time series instances. Set by algorithm.

targets

protected int[] targets

Target values for the instances. Set by the algorithm.

nearestNeighborClassifier

protected NearestNeighborClassifier nearestNeighborClassifier

The nearest neighbor classifier used for prediction. Set by the algorithm.

shotgunDistance

protected ShotgunDistance shotgunDistance

The Shotgun Distance used by the nearestNeighborClassifier. Set by the algorithm.

windows

protected java.util.ArrayList<ai.libs.jaicore.basic.sets.Pair<java.lang.Integer,java.lang.Integer>> windows

Holds pairs of (number of correct predictions, window length) obtained in training phase.

bestScore

protected int bestScore

The best score. States the highest number of correct predicitions for every window length used in training phase (leave-one-out).

Constructor Detail

ShotgunEnsembleClassifier

public ShotgunEnsembleClassifier(int minWindowLength,
                                 int maxWindowLength,
                                 boolean meanNormalization,
                                 double factor)

Creates a Shotgun Ensemble classifier.

Parameters:

algorithm - The training algorithm.

factor - Factor used to determine whether or not to include a window length into the overall predicition.

Method Detail

calculateWindowLengthPredictions

protected java.util.Map<java.lang.Integer,java.lang.Integer> calculateWindowLengthPredictions(double[] testInstance)
                                                                                       throws PredictionException

Calculates predicitions for a test instance using 1NN with Shotgun Distance and different window lengths.

Parameters:

testInstance - The test instance.

Returns:

Map of (window length, prediciton) pairs.

Throws:

PredictionException

mostFrequentLabelFromWindowLengthPredicitions

protected java.lang.Integer mostFrequentLabelFromWindowLengthPredicitions(java.util.Map<java.lang.Integer,java.lang.Integer> windowLengthPredicitions)

Returns the most frequent predicition given a Map of (window length, prediciton) pairs.

Parameters:

windowLengthPredicitions - Map of (window length, prediciton) pairs.

Returns:

The most frequent predicition.

calculateWindowLengthPredictions

protected java.util.Map<java.lang.Integer,java.util.List<java.lang.Integer>> calculateWindowLengthPredictions(TimeSeriesDataset dataset)
                                                                                                       throws PredictionException

Calculates predicitions for a test dataset using 1NN with Shotgun Distance and different window lengths.

Parameters:

dataset - The dataset to predict for.

Returns:

Map of (window length, predicitions) pairs.

Throws:

PredictionException

mostFrequentLabelsFromWindowLengthPredicitions

protected java.util.List<java.lang.Integer> mostFrequentLabelsFromWindowLengthPredicitions(java.util.Map<java.lang.Integer,java.util.List<java.lang.Integer>> windowLengthPredicitions)

Returns for each instance the most frequent predicitions as contained in a Map of (window length, list of prediciton for each instance) pairs.

Parameters:

windowLengthPredicitions - Map of (window length, list of prediciton for each instance) pairs.

Returns:

The most frequent predicition for each instace.

predict

public java.lang.Integer predict(double[] univInstance)
                          throws PredictionException

Predicts on univariate instance.

Specified by:

predict in class ASimplifiedTSClassifier<java.lang.Integer>

Parameters:

univInstance - The univariate instance.

Returns:

Class prediction for the instance.

Throws:

PredictionException - If something fails during the prediction process.

predict

public java.lang.Integer predict(java.util.List<double[]> multivInstance)
                          throws PredictionException

Predicts on a multivariate instance. This is not supported yet.

Specified by:

predict in class ASimplifiedTSClassifier<java.lang.Integer>

Parameters:

multivInstance - The multivariate instance.

Returns:

Class prediciton for the instance.

Throws:

PredictionException - If something fails during the prediction process.

predict

public java.util.List<java.lang.Integer> predict(TimeSeriesDataset dataset)
                                          throws PredictionException

Predicts on a dataset.

Specified by:

predict in class ASimplifiedTSClassifier<java.lang.Integer>

Parameters:

dataset - The dataset.

Returns:

List of class predicitons for each instance of the dataset.

Throws:

PredictionException - If something fails during the prediction process

setValues

protected void setValues(double[][] values)

Sets the value matrix.

Parameters:

values -

setTargets

protected void setTargets(int[] targets)

Sets the targets.

Parameters:

targets -

setWindows

protected void setWindows(java.util.ArrayList<ai.libs.jaicore.basic.sets.Pair<java.lang.Integer,java.lang.Integer>> windows)

Sets the windows and also retreives and sets the @see #bestScore from these windows.

Parameters:

windows - @see #windows

setNearestNeighborClassifier

protected void setNearestNeighborClassifier(NearestNeighborClassifier nearestNeighborClassifier)

Sets the nearest neighbor classifier, nearestNeighborClassifier.

Parameters:

nearestNeighborClassifier -

getLearningAlgorithm

public ShotgunEnsembleLearnerAlgorithm getLearningAlgorithm(TimeSeriesDataset dataset)

Specified by:

getLearningAlgorithm in class ASimplifiedTSClassifier<java.lang.Integer>