SpearmansCorrelation.java

/*
 * Licensed to the Apache Software Foundation (ASF) under one or more
 * contributor license agreements.  See the NOTICE file distributed with
 * this work for additional information regarding copyright ownership.
 * The ASF licenses this file to You under the Apache License, Version 2.0
 * (the "License"); you may not use this file except in compliance with
 * the License.  You may obtain a copy of the License at
 *
 *      http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

package org.apache.commons.math4.legacy.stat.correlation;

import org.apache.commons.math4.legacy.exception.DimensionMismatchException;
import org.apache.commons.math4.legacy.exception.MathIllegalArgumentException;
import org.apache.commons.math4.legacy.exception.util.LocalizedFormats;
import org.apache.commons.math4.legacy.linear.BlockRealMatrix;
import org.apache.commons.math4.legacy.linear.RealMatrix;
import org.apache.commons.math4.legacy.stat.ranking.NaNStrategy;
import org.apache.commons.math4.legacy.stat.ranking.NaturalRanking;
import org.apache.commons.math4.legacy.stat.ranking.RankingAlgorithm;

/**
 * Spearman's rank correlation. This implementation performs a rank
 * transformation on the input data and then computes {@link PearsonsCorrelation}
 * on the ranked data.
 * <p>
 * By default, ranks are computed using {@link NaturalRanking} with default
 * strategies for handling NaNs and ties in the data (NaNs maximal, ties averaged).
 * The ranking algorithm can be set using a constructor argument.
 *
 * @since 2.0
 */
public class SpearmansCorrelation {

    /** Input data. */
    private final RealMatrix data;

    /** Ranking algorithm.  */
    private final RankingAlgorithm rankingAlgorithm;

    /** Rank correlation. */
    private final PearsonsCorrelation rankCorrelation;

    /**
     * Create a SpearmansCorrelation without data.
     */
    public SpearmansCorrelation() {
        this(new NaturalRanking());
    }

    /**
     * Create a SpearmansCorrelation with the given ranking algorithm.
     *
     * @param rankingAlgorithm ranking algorithm
     * @throws MathIllegalArgumentException if the provided {@link RankingAlgorithm} is of
     * type {@link NaturalRanking} and uses a {@link NaNStrategy#REMOVED} strategy
     * @since 3.1
     */
    public SpearmansCorrelation(final RankingAlgorithm rankingAlgorithm)
        throws MathIllegalArgumentException {

        if (rankingAlgorithm instanceof NaturalRanking &&
            NaNStrategy.REMOVED == ((NaturalRanking) rankingAlgorithm).getNanStrategy()) {
            throw new MathIllegalArgumentException(LocalizedFormats.NOT_SUPPORTED_NAN_STRATEGY,
                                                   NaNStrategy.REMOVED);
        }

        data = null;
        this.rankingAlgorithm = rankingAlgorithm;
        rankCorrelation = null;
    }

    /**
     * Create a SpearmansCorrelation from the given data matrix.
     *
     * @param dataMatrix matrix of data with columns representing
     * variables to correlate
     */
    public SpearmansCorrelation(final RealMatrix dataMatrix) {
        this(dataMatrix, new NaturalRanking());
    }

    /**
     * Create a SpearmansCorrelation with the given input data matrix
     * and ranking algorithm.
     *
     * @param dataMatrix matrix of data with columns representing
     * variables to correlate
     * @param rankingAlgorithm ranking algorithm
     * @throws MathIllegalArgumentException if the provided {@link RankingAlgorithm} is of
     * type {@link NaturalRanking} and uses a {@link NaNStrategy#REMOVED} strategy
     */
    public SpearmansCorrelation(final RealMatrix dataMatrix, final RankingAlgorithm rankingAlgorithm)
        throws MathIllegalArgumentException {

        if (rankingAlgorithm instanceof NaturalRanking &&
            NaNStrategy.REMOVED == ((NaturalRanking) rankingAlgorithm).getNanStrategy()) {
            throw new MathIllegalArgumentException(LocalizedFormats.NOT_SUPPORTED_NAN_STRATEGY,
                                                   NaNStrategy.REMOVED);
        }

        this.rankingAlgorithm = rankingAlgorithm;
        this.data = rankTransform(dataMatrix);
        rankCorrelation = new PearsonsCorrelation(data);
    }

    /**
     * Calculate the Spearman Rank Correlation Matrix.
     *
     * @return Spearman Rank Correlation Matrix
     * @throws NullPointerException if this instance was created with no data.
     */
    public RealMatrix getCorrelationMatrix() {
        return rankCorrelation.getCorrelationMatrix();
    }

    /**
     * Returns a {@link PearsonsCorrelation} instance constructed from the
     * ranked input data. That is,
     * <code>new SpearmansCorrelation(matrix).getRankCorrelation()</code>
     * is equivalent to
     * <code>new PearsonsCorrelation(rankTransform(matrix))</code> where
     * <code>rankTransform(matrix)</code> is the result of applying the
     * configured <code>RankingAlgorithm</code> to each of the columns of
     * <code>matrix.</code>
     *
     * <p>Returns null if this instance was created with no data.</p>
     *
     * @return PearsonsCorrelation among ranked column data
     */
    public PearsonsCorrelation getRankCorrelation() {
        return rankCorrelation;
    }

    /**
     * Computes the Spearman's rank correlation matrix for the columns of the
     * input matrix.
     *
     * @param matrix matrix with columns representing variables to correlate
     * @return correlation matrix
     */
    public RealMatrix computeCorrelationMatrix(final RealMatrix matrix) {
        final RealMatrix matrixCopy = rankTransform(matrix);
        return new PearsonsCorrelation().computeCorrelationMatrix(matrixCopy);
    }

    /**
     * Computes the Spearman's rank correlation matrix for the columns of the
     * input rectangular array.  The columns of the array represent values
     * of variables to be correlated.
     *
     * @param matrix matrix with columns representing variables to correlate
     * @return correlation matrix
     */
    public RealMatrix computeCorrelationMatrix(final double[][] matrix) {
       return computeCorrelationMatrix(new BlockRealMatrix(matrix));
    }

    /**
     * Computes the Spearman's rank correlation coefficient between the two arrays.
     *
     * @param xArray first data array
     * @param yArray second data array
     * @return Returns Spearman's rank correlation coefficient for the two arrays
     * @throws DimensionMismatchException if the arrays lengths do not match
     * @throws MathIllegalArgumentException if the array length is less than 2
     */
    public double correlation(final double[] xArray, final double[] yArray) {
        if (xArray.length != yArray.length) {
            throw new DimensionMismatchException(xArray.length, yArray.length);
        } else if (xArray.length < 2) {
            throw new MathIllegalArgumentException(LocalizedFormats.INSUFFICIENT_DIMENSION,
                                                   xArray.length, 2);
        } else {
            return new PearsonsCorrelation().correlation(rankingAlgorithm.rank(xArray),
                                                         rankingAlgorithm.rank(yArray));
        }
    }

    /**
     * Applies rank transform to each of the columns of <code>matrix</code>
     * using the current <code>rankingAlgorithm</code>.
     *
     * @param matrix matrix to transform
     * @return a rank-transformed matrix
     */
    private RealMatrix rankTransform(final RealMatrix matrix) {
        RealMatrix transformed = matrix.copy();
        for (int i = 0; i < transformed.getColumnDimension(); i++) {
            transformed.setColumn(i, rankingAlgorithm.rank(transformed.getColumn(i)));
        }

        return transformed;
    }
}