org.openimaj.ml.clustering.assignment.hard.HierarchicalByteHardAssigner 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/ml/clustering/assignment/hard/Hierarchical#T#HardAssigner.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.ml.clustering.assignment.hard;

import org.openimaj.ml.clustering.assignment.HardAssigner;
import org.openimaj.ml.clustering.assignment.soft.HierarchicalBytePathAssigner;
import org.openimaj.ml.clustering.kmeans.HierarchicalByteKMeansResult;
import org.openimaj.util.pair.IndependentPair;
import org.openimaj.util.pair.IntFloatPair;

/**
 * The {@link HierarchicalByteHardAssigner} is a {@link HardAssigner} for
 * {@link HierarchicalByteKMeansResult} instances. The assigner
 * produces the index of the assigned leaf node as if the clusters were
 * actually flat. 
 * 
 * @author Jonathon Hare ([email protected])
 */
public class HierarchicalByteHardAssigner implements HardAssigner {
	/**
	 * The {@link ScoringScheme} determines how the distance
	 * to a cluster is estimated from the hierarchy of k-means
	 * generated clusters.
	 * 
	 * @author Jonathon Hare ([email protected])
	 */
	public enum ScoringScheme {
		/**
		 * Sum distances down the tree.
		 * 
		 * @author Jonathon Hare ([email protected])
		 */
		SUM {
			@Override
			public float computeScore(float[] weights) {
				float sum = 0;
				for (float w : weights) {
					if (w < 0) break;
					sum += w;
				}
				
				return sum;
			}
		},
		/**
		 * Product of distances down the tree.
		 * 
		 * @author Jonathon Hare ([email protected])
		 */
		PRODUCT {
			@Override
			public float computeScore(float[] weights) {
				float prod = 1;
				for (float w : weights) {
					if (w < 0) break;
					prod *= w;
				}
				
				return prod;
			}
		},
		/**
		 * The distance in the root cluster 
		 * 
		 * @author Jonathon Hare ([email protected])
		 */
		FIRST {
			@Override
			public float computeScore(float[] weights) {
				return weights[0];
			}
		},
		/**
		 * The distance in the leaf cluster
		 * 
		 * @author Jonathon Hare ([email protected])
		 */
		LAST {
			@Override
			public float computeScore(float[] weights) {
				float last = -1;
				
				for (float w : weights) {
					if (w < 0) break;
					last = w;
				}
				
				return last;
			}
		},
		/**
		 * The mean distance down the tree
		 * 
		 * @author Jonathon Hare ([email protected])
		 */
		MEAN {
			@Override
			public float computeScore(float[] weights) {
				float sum = 0;
				int count = 0;
				
				for (float w : weights) {
					if (w < 0) break;
					sum += w;
					count++;
				}
				
				return sum / (float)count;
			}
		}
		;
		
		protected abstract float computeScore(float[] weights); 
	}
	
	protected HierarchicalByteKMeansResult result;
	protected HierarchicalBytePathAssigner path;
	protected ScoringScheme scorer;
	
	/**
	 * Construct with the given hierarchical KMeans clusterer
	 * and scoring scheme.
	 *
	 * @param result the hierarchical KMeans clusterer
	 * @param scorer the scoring scheme
	 */
	public HierarchicalByteHardAssigner(HierarchicalByteKMeansResult result, ScoringScheme scorer) {
		this.result = result;
		this.scorer = scorer;
		this.path = new HierarchicalBytePathAssigner(result);
	}

	/**
	 * Construct with the given Hierarchical KMeans clusterer
	 * and the SUM scoring scheme.
	 *
	 * @param result the hierarchical KMeans clusterer
	 */
	public HierarchicalByteHardAssigner(HierarchicalByteKMeansResult result) {
		this(result, ScoringScheme.SUM);
	}
	
	@Override
	public int[] assign(byte[][] data) {
		int [] asgn = new int[data.length];

		for (int i=0; i pw = path.assignWeighted(data);
		
		int index = result.getIndex(pw.firstObject());
		float score = scorer.computeScore(pw.secondObject());
		
		return new IntFloatPair(index, score);
	}
	
	@Override
	public int size() {
	    return result.countLeafs();
	}
	
	@Override
	public int numDimensions() {
	    return result.numDimensions();
	}
}