org.openimaj.knn.pq.LongProductQuantiserUtilities Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of clustering Show documentation
Various clustering algorithm implementations for all primitive types including random, random forest, K-Means (Exact, Hierarchical and Approximate), ...
There is a newer version: 1.3.10
Show newest version
/*
	AUTOMATICALLY GENERATED BY jTemp FROM
	/Users/jsh2/Work/openimaj/target/checkout/machine-learning/clustering/src/main/jtemp/org/openimaj/knn/pq/#T#ProductQuantiserUtilities.jtemp
*/
/**
 * Copyright (c) 2011, The University of Southampton and the individual contributors.
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without modification,
 * are permitted provided that the following conditions are met:
 *
 *   * 	Redistributions of source code must retain the above copyright notice,
 * 	this list of conditions and the following disclaimer.
 *
 *   *	Redistributions in binary form must reproduce the above copyright notice,
 * 	this list of conditions and the following disclaimer in the documentation
 * 	and/or other materials provided with the distribution.
 *
 *   *	Neither the name of the University of Southampton nor the names of its
 * 	contributors may be used to endorse or promote products derived from this
 * 	software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
 * ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */
package org.openimaj.knn.pq;

import org.openimaj.knn.LongNearestNeighboursExact;
import org.openimaj.knn.LongNearestNeighboursProvider;
import org.openimaj.ml.clustering.kmeans.LongKMeans;

/**
 * Utility methods for easily creating a {@link ByteProductQuantiser} using
 * (Exact) K-Means.
 * 
 * @author Jonathon Hare ([email protected])
 * 
 */
public final class LongProductQuantiserUtilities {
    private LongProductQuantiserUtilities() {
    }

	/**
	 * Learn a {@link LongProductQuantiser} by applying exact K-Means to
	 * sub-vectors extracted from the given data. The length of the subvectors
	 * is determined by dividing the vector length by the requested number of
	 * assigners. The number of clusters per vector subset is constant, and must
	 * be less than 256.
	 * 
	 * @param data
	 *            the data to train the {@link LongProductQuantiser} on.
	 * @param numAssigners
	 *            the number of sub-quantisers to learn
	 * @param K
	 *            the number of centroids per sub-quantiser
	 * @param nIter
	 *            the maximum number of iterations for each k-means clustering	 
	 * 
	 * @return a trained {@link LongProductQuantiser}.
	 */
	public static LongProductQuantiser train(long[][] data, int numAssigners, int K, int nIter) {
		if (K > 256 || K <= 0)
			throw new IllegalArgumentException("0 <= K < 256");

		final int subDim = data[0].length / numAssigners;
		final long[][] tmp = new long[data.length][subDim];

		final LongNearestNeighboursExact[] assigners = new LongNearestNeighboursExact[numAssigners];
		final LongKMeans kmeans = LongKMeans.createExact(K, 100);

		for (int i = 0; i < numAssigners; i++) {
			// copy data
			for (int j = 0; j < data.length; j++) {
				System.arraycopy(data[j], i * subDim, tmp[j], 0, subDim);
			}

			// kmeans
			final LongNearestNeighboursProvider centroids = (LongNearestNeighboursProvider) kmeans.cluster(tmp);

			assigners[i] = (LongNearestNeighboursExact)centroids.getNearestNeighbours();
		}

		return new LongProductQuantiser(assigners);
	}

	/**
	 * Learn a {@link LongProductQuantiser} by applying exact K-Means to
	 * sub-vectors extracted from the given data. The length of the subvectors
	 * is determined by dividing the vector length by the requested number of
	 * assigners. The number of clusters per vector subset is constant, and set
	 * at 256.
	 * 
	 * @param data
	 *            the data to train the {@link LongProductQuantiser} on.
	 * @param numAssigners
	 *            the number of sub-quantisers to learn
	 * @param nIter
	 *            the maximum number of iterations for each k-means clustering
	 *
	 * @return a trained {@link LongProductQuantiser}.
	 */
	public static LongProductQuantiser train(long[][] data, int numAssigners, int nIter) {
		return train(data, numAssigners, 256, nIter);
	}
}