You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@commons.apache.org by ps...@apache.org on 2013/11/09 22:32:06 UTC
svn commit: r1540395 -
/commons/proper/math/trunk/src/main/java/org/apache/commons/math3/stat/correlation/PearsonsCorrelation.java
Author: psteitz
Date: Sat Nov 9 21:32:06 2013
New Revision: 1540395
URL: http://svn.apache.org/r1540395
Log:
Clarified contracts re NaNs, IAEs and when constructor arguments are necessary.
Modified:
commons/proper/math/trunk/src/main/java/org/apache/commons/math3/stat/correlation/PearsonsCorrelation.java
Modified: commons/proper/math/trunk/src/main/java/org/apache/commons/math3/stat/correlation/PearsonsCorrelation.java
URL: http://svn.apache.org/viewvc/commons/proper/math/trunk/src/main/java/org/apache/commons/math3/stat/correlation/PearsonsCorrelation.java?rev=1540395&r1=1540394&r2=1540395&view=diff
==============================================================================
--- commons/proper/math/trunk/src/main/java/org/apache/commons/math3/stat/correlation/PearsonsCorrelation.java (original)
+++ commons/proper/math/trunk/src/main/java/org/apache/commons/math3/stat/correlation/PearsonsCorrelation.java Sat Nov 9 21:32:06 2013
@@ -34,9 +34,17 @@ import org.apache.commons.math3.util.Fas
* <code>double[][]</code> arguments generate correlation matrices. The
* columns of the input matrices are assumed to represent variable values.
* Correlations are given by the formula</p>
- * <code>cor(X, Y) = Σ[(x<sub>i</sub> - E(X))(y<sub>i</sub> - E(Y))] / [(n - 1)s(X)s(Y)]</code>
+ *
+ * <p><code>cor(X, Y) = Σ[(x<sub>i</sub> - E(X))(y<sub>i</sub> - E(Y))] / [(n - 1)s(X)s(Y)]</code>
* where <code>E(X)</code> is the mean of <code>X</code>, <code>E(Y)</code>
- * is the mean of the <code>Y</code> values and s(X), s(Y) are standard deviations.
+ * is the mean of the <code>Y</code> values and s(X), s(Y) are standard deviations.</p>
+ *
+ * <p>To compute the correlation coefficient for a single pair of arrays, use {@link #PearsonsCorrelation()}
+ * to construct an instance with no data and then {@link #correlation(double[], double[])}.
+ * Correlation matrices can also be computed directly from an instance with no data using
+ * {@link #computeCorrelationMatrix(double[][])}. In order to use {@link #getCorrelationMatrix()},
+ * {@link #getCorrelationPValues()}, or {@link #getCorrelationStandardErrors()}; however, one of the
+ * constructors supplying data or a covariance matrix must be used to create the instance.</p>
*
* @version $Id$
* @since 2.0
@@ -50,7 +58,7 @@ public class PearsonsCorrelation {
private final int nObs;
/**
- * Create a PearsonsCorrelation instance without data
+ * Create a PearsonsCorrelation instance without data.
*/
public PearsonsCorrelation() {
super();
@@ -62,9 +70,14 @@ public class PearsonsCorrelation {
* Create a PearsonsCorrelation from a rectangular array
* whose columns represent values of variables to be correlated.
*
+ * Throws MathIllegalArgumentException if the input array does not have at least
+ * two columns and two rows. Pairwise correlations are set to NaN if one
+ * of the correlates has zero variance.
+ *
* @param data rectangular array with columns representing variables
- * @throws IllegalArgumentException if the input data array is not
+ * @throws MathIllegalArgumentException if the input data array is not
* rectangular with at least two rows and two columns.
+ * @see #correlation(double[], double[])
*/
public PearsonsCorrelation(double[][] data) {
this(new BlockRealMatrix(data));
@@ -74,10 +87,15 @@ public class PearsonsCorrelation {
* Create a PearsonsCorrelation from a RealMatrix whose columns
* represent variables to be correlated.
*
+ * Throws MathIllegalArgumentException if the matrix does not have at least
+ * two columns and two rows. Pairwise correlations are set to NaN if one
+ * of the correlates has zero variance.
+ *
* @param matrix matrix with columns representing variables to correlate
+ * @throws MathIllegalArgumentException if the matrix does not contain sufficient data
+ * @see #correlation(double[], double[])
*/
public PearsonsCorrelation(RealMatrix matrix) {
- checkSufficientData(matrix);
nObs = matrix.getRowDimension();
correlationMatrix = computeCorrelationMatrix(matrix);
}
@@ -100,7 +118,7 @@ public class PearsonsCorrelation {
}
/**
- * Create a PearsonsCorrelation from a covariance matrix. The correlation
+ * Create a PearsonsCorrelation from a covariance matrix. The correlation
* matrix is computed by scaling the covariance matrix.
*
* @param covarianceMatrix covariance matrix
@@ -110,11 +128,14 @@ public class PearsonsCorrelation {
public PearsonsCorrelation(RealMatrix covarianceMatrix, int numberOfObservations) {
nObs = numberOfObservations;
correlationMatrix = covarianceToCorrelation(covarianceMatrix);
-
}
/**
- * Returns the correlation matrix
+ * Returns the correlation matrix.
+ *
+ * <p>This method will return null if the argumentless constructor was used
+ * to create this instance, even if {@link #computeCorrelationMatrix(double[][])}
+ * has been called before it is activated.</p>
*
* @return correlation matrix
*/
@@ -127,12 +148,17 @@ public class PearsonsCorrelation {
* in the correlation matrix.<br/>
* <code>getCorrelationStandardErrors().getEntry(i,j)</code> is the standard
* error associated with <code>getCorrelationMatrix.getEntry(i,j)</code>
+ *
* <p>The formula used to compute the standard error is <br/>
* <code>SE<sub>r</sub> = ((1 - r<sup>2</sup>) / (n - 2))<sup>1/2</sup></code>
* where <code>r</code> is the estimated correlation coefficient and
* <code>n</code> is the number of observations in the source dataset.</p>
*
+ * <p>To use this method, one of the constructors that supply an input
+ * matrix must have been used to create this instance.</p>
+ *
* @return matrix of correlation standard errors
+ * @throws NullPointerException if this instance was created with no data
*/
public RealMatrix getCorrelationStandardErrors() {
int nVars = correlationMatrix.getColumnDimension();
@@ -149,16 +175,22 @@ public class PearsonsCorrelation {
/**
* Returns a matrix of p-values associated with the (two-sided) null
* hypothesis that the corresponding correlation coefficient is zero.
+ *
* <p><code>getCorrelationPValues().getEntry(i,j)</code> is the probability
* that a random variable distributed as <code>t<sub>n-2</sub></code> takes
* a value with absolute value greater than or equal to <br>
* <code>|r|((n - 2) / (1 - r<sup>2</sup>))<sup>1/2</sup></code></p>
+ *
* <p>The values in the matrix are sometimes referred to as the
* <i>significance</i> of the corresponding correlation coefficients.</p>
*
+ * <p>To use this method, one of the constructors that supply an input
+ * matrix must have been used to create this instance.</p>
+ *
* @return matrix of p-values
* @throws org.apache.commons.math3.exception.MaxCountExceededException
* if an error occurs estimating probabilities
+ * @throws NullPointerException if this instance was created with no data
*/
public RealMatrix getCorrelationPValues() {
TDistribution tDistribution = new TDistribution(nObs - 2);
@@ -181,12 +213,19 @@ public class PearsonsCorrelation {
/**
* Computes the correlation matrix for the columns of the
- * input matrix.
+ * input matrix, using {@link #correlation(double[], double[])}.
+ *
+ * Throws MathIllegalArgumentException if the matrix does not have at least
+ * two columns and two rows. Pairwise correlations are set to NaN if one
+ * of the correlates has zero variance.
*
* @param matrix matrix with columns representing variables to correlate
* @return correlation matrix
+ * @throws MathIllegalArgumentException if the matrix does not contain sufficient data
+ * @see #correlation(double[], double[])
*/
public RealMatrix computeCorrelationMatrix(RealMatrix matrix) {
+ checkSufficientData(matrix);
int nVars = matrix.getColumnDimension();
RealMatrix outMatrix = new BlockRealMatrix(nVars, nVars);
for (int i = 0; i < nVars; i++) {
@@ -202,21 +241,29 @@ public class PearsonsCorrelation {
/**
* Computes the correlation matrix for the columns of the
- * input rectangular array. The colums of the array represent values
+ * input rectangular array. The columns of the array represent values
* of variables to be correlated.
*
+ * Throws MathIllegalArgumentException if the matrix does not have at least
+ * two columns and two rows or if the array is not rectangular. Pairwise
+ * correlations are set to NaN if one of the correlates has zero variance.
+ *
* @param data matrix with columns representing variables to correlate
* @return correlation matrix
+ * @throws MathIllegalArgumentException if the array does not contain sufficient data
+ * @see #correlation(double[], double[])
*/
public RealMatrix computeCorrelationMatrix(double[][] data) {
return computeCorrelationMatrix(new BlockRealMatrix(data));
}
/**
- * Computes the Pearson's product-moment correlation coefficient between the two arrays.
+ * Computes the Pearson's product-moment correlation coefficient between two arrays.
*
- * </p>Throws IllegalArgumentException if the arrays do not have the same length
- * or their common length is less than 2</p>
+ * <p>Throws MathIllegalArgumentException if the arrays do not have the same length
+ * or their common length is less than 2. Returns {@code NaN} if either of the arrays
+ * has zero variance (i.e., if one of the arrays does not contain at least two distinct
+ * values).</p>
*
* @param xArray first data array
* @param yArray second data array
@@ -267,8 +314,8 @@ public class PearsonsCorrelation {
}
/**
- * Throws IllegalArgumentException of the matrix does not have at least
- * two columns and two rows
+ * Throws MathIllegalArgumentException if the matrix does not have at least
+ * two columns and two rows.
*
* @param matrix matrix to check for sufficiency
* @throws MathIllegalArgumentException if there is insufficient data