Added implementation of differential entropy for Gaussian variables.

java/source/infodynamics/measures/continuous/kernel/TransferEntropyCalculatorKernel.java

@@ -61,7 +61,7 @@ public class TransferEntropyCalculatorKernel
 	public static final String FORCE_KERNEL_COMPARE_TO_ALL = "FORCE_KERNEL_COMPARE_TO_ALL";
 	
 	/**
-	 * Default value for epsilon
+	 * Default value for epsilon (kernel width)
 	 */
 	public static final double DEFAULT_EPSILON = 0.25;
 	/**

java/source/infodynamics/measures/continuous/kraskov/TransferEntropyCalculatorKraskovByMulti.java

@@ -51,7 +51,7 @@ import java.util.Iterator;
  * <ol>
  * 		<li>Construct</li>
  * 		<li>SetProperty() for each property</li>
- * 		<li>intialise()</li>
+ * 		<li>initialise()</li>
  * 		<li>setObservations(), or [startAddObservations(), addObservations()*, finaliseAddObservations()]
  *   Note: If not using setObservations(), the results from computeLocal or getSignificance
  *    are not likely to be particularly sensible.</li> 
@@ -99,6 +99,9 @@ public class TransferEntropyCalculatorKraskovByMulti
 		createKraskovMiCalculators();
 	}
 
+	/**
+	 * @param k history length (Schreiber k parameter, not Kraskov k parameter)
+	 */
 	public void initialise(int k) throws Exception {
 		super.initialise(k); // calls initialise();
 	}

java/source/infodynamics/measures/continuous/lineargaussian/EntropyCalculatorLinearGaussian.java

@@ -0,0 +1,90 @@
+package infodynamics.measures.continuous.lineargaussian;
+
+import infodynamics.measures.continuous.EntropyCalculator;
+import infodynamics.utils.MatrixUtils;
+
+/**
+ * <p>Computes the differential entropy of a given set of observations, assuming that
+ *  the probability distribution function for these observations is Gaussian.</p>
+ *  
+ * <p>
+ * Usage:
+ * 	<ol>
+ * 		<li>Construct</li>
+ *		<li>initialise()</li>
+ * 		<li>setObservations(), or setVariance().</li> 
+ * 		<li>computeAverageLocalOfObservations() to return the average differential
+ *          entropy based on either the set variance or the variance of
+ *          the supplied observations.</li>
+ * 	</ol>
+ * </p>
+ * 
+ * @author Joseph Lizier joseph.lizier_at_gmail.com
+ *
+ */
+public class EntropyCalculatorLinearGaussian implements EntropyCalculator {
+
+	/**
+	 * Variance of the most recently supplied observations
+	 */
+	protected double variance;
+	
+	protected boolean debug;
+	
+	/**
+	 * Constructor
+	 */
+	public EntropyCalculatorLinearGaussian() {
+		// Nothing to do
+	}
+	
+	/**
+	 * Initialise the calculator ready for reuse
+	 */
+	public void initialise() throws Exception {
+		// Nothing to do
+	}
+
+	/**
+	 * Provide the observations from which to compute the entropy
+	 * 
+	 * @param observations the observations to compute the entropy from
+	 */
+	public void setObservations(double[] observations) {
+		variance = MatrixUtils.stdDev(observations);
+		variance *= variance;
+	}
+
+	/**
+	 * Set the variance of the distribution for which we will compute the
+	 *  entropy.
+	 * 
+	 * @param variance
+	 */
+	public void setVariance(double variance) {
+		this.variance = variance;
+	}
+	
+	/**
+	 * The entropy for a Gaussian-distribution random variable with
+	 *  variance \sigma is \log_e{2*pi*e*\sigma}.
+	 * Here we compute the entropy assuming that the recorded estimation of the
+	 *  variance is correct (i.e. we will not make a bias correction for limited
+	 *  observations here).
+	 * 
+	 * @return the entropy of the previously provided observations
+	 */
+	public double computeAverageLocalOfObservations() {
+		return Math.log(2.0*Math.PI*Math.E*variance);
+	}
+
+	public void setDebug(boolean debug) {
+		this.debug = debug;
+	}
+
+	public void setProperty(String propertyName, String propertyValue)
+			throws Exception {
+		// No properties to set here
+	}
+
+}

java/source/infodynamics/measures/discrete/ApparentTransferEntropyCalculator.java

@@ -17,6 +17,9 @@ import infodynamics.utils.RandomGenerator;
  * 2. Standalone computation from a single set of observations:
  *   Call: computeLocal() or computeAverageLocal()
  * 
+ * @see For transfer entropy: Schreiber, PRL 85 (2) pp.461-464, 2000; http://dx.doi.org/10.1103/PhysRevLett.85.461
+ * @see For local transfer entropy: Lizier et al, PRE 77, 026110, 2008; http://dx.doi.org/10.1103/PhysRevE.77.026110
+ * 
  * @author Joseph Lizier
  * joseph.lizier at gmail.com
  * http://lizier.me/joseph/
@@ -58,6 +61,14 @@ public class ApparentTransferEntropyCalculator extends ContextOfPastMeasureCalcu
 		*/
 	}
 	
+	/**
+	 * Create a new TE calculator for the given base and history length.
+	 * 
+	 * @param base number of quantisation levels for each variable.
+	 *        E.g. binary variables are in base-2.
+	 * @param history history length of the destination to condition on -
+	 *        this is k in Schreiber's notation.
+	 */
 	public ApparentTransferEntropyCalculator(int base, int history) {
 
 		super(base, history);

java/source/infodynamics/utils/MatrixUtils.java

@@ -2544,8 +2544,8 @@ public class MatrixUtils {
 	
 	/**
 	 * <p>Private function to compute the determinant recursively.
-	 * determinant() calls this after checking the matrix dimensions. <br/>
-	 * See - http://mathworld.wolfram.com/Determinant.html
+	 * {@link determinant()} calls this after checking the matrix dimensions. <br/>
+	 * @see {@link http://mathworld.wolfram.com/Determinant.html}
 	 * </p>
 	 * 
 	 * @param matrix