bayesnet.java

来自「Weka」· Java 代码 · 共 1,118 行 · 第 1/3 页

	    StringBuffer text = new StringBuffer();
	    text.append("<?xml version=\"1.0\"?>\n");
	    text.append("<!-- DTD for the XMLBIF 0.3 format -->\n");
	    text.append("<!DOCTYPE BIF [\n");
	    text.append("	<!ELEMENT BIF ( NETWORK )*>\n");
	    text.append("	      <!ATTLIST BIF VERSION CDATA #REQUIRED>\n");
	    text.append("	<!ELEMENT NETWORK ( NAME, ( PROPERTY | VARIABLE | DEFINITION )* )>\n");
	    text.append("	<!ELEMENT NAME (#PCDATA)>\n");
	    text.append("	<!ELEMENT VARIABLE ( NAME, ( OUTCOME |  PROPERTY )* ) >\n");
	    text.append("	      <!ATTLIST VARIABLE TYPE (nature|decision|utility) \"nature\">\n");
	    text.append("	<!ELEMENT OUTCOME (#PCDATA)>\n");
	    text.append("	<!ELEMENT DEFINITION ( FOR | GIVEN | TABLE | PROPERTY )* >\n");
	    text.append("	<!ELEMENT FOR (#PCDATA)>\n");
	    text.append("	<!ELEMENT GIVEN (#PCDATA)>\n");
	    text.append("	<!ELEMENT TABLE (#PCDATA)>\n");
	    text.append("	<!ELEMENT PROPERTY (#PCDATA)>\n");
	    text.append("]>\n");
	    return text.toString();
  } // getBIFHeader
  
  /**
   * Returns a description of the classifier in XML BIF 0.3 format.
   * See http://www-2.cs.cmu.edu/~fgcozman/Research/InterchangeFormat/
   * for details on XML BIF.
   * @return an XML BIF 0.3 description of the classifier as a string.
   */
  public String toXMLBIF03() {
    if (m_Instances == null) {
      return("<!--No model built yet-->");
    }
    
    StringBuffer text = new StringBuffer();
    text.append(getBIFHeader());
    text.append("\n");
    text.append("\n");
    text.append("<BIF VERSION=\"0.3\">\n");
    text.append("<NETWORK>\n");
    text.append("<NAME>" + XMLNormalize(m_Instances.relationName()) + "</NAME>\n");
    for (int iAttribute = 0; iAttribute < m_Instances.numAttributes(); iAttribute++) {
      text.append("<VARIABLE TYPE=\"nature\">\n");
      text.append("<NAME>" + XMLNormalize(m_Instances.attribute(iAttribute).name()) + "</NAME>\n");
      for (int iValue = 0; iValue < m_Instances.attribute(iAttribute).numValues(); iValue++) {
        text.append("<OUTCOME>" + XMLNormalize(m_Instances.attribute(iAttribute).value(iValue)) + "</OUTCOME>\n");
      }
      text.append("</VARIABLE>\n");
    }
    
    for (int iAttribute = 0; iAttribute < m_Instances.numAttributes(); iAttribute++) {
      text.append("<DEFINITION>\n");
      text.append("<FOR>" + XMLNormalize(m_Instances.attribute(iAttribute).name()) + "</FOR>\n");
      for (int iParent = 0; iParent < m_ParentSets[iAttribute].getNrOfParents(); iParent++) {
        text.append("<GIVEN>"
        + XMLNormalize(m_Instances.attribute(m_ParentSets[iAttribute].getParent(iParent)).name()) +
        "</GIVEN>\n");
      }
      text.append("<TABLE>\n");
      for (int iParent = 0; iParent < m_ParentSets[iAttribute].getCardinalityOfParents(); iParent++) {
        for (int iValue = 0; iValue < m_Instances.attribute(iAttribute).numValues(); iValue++) {
          text.append(m_Distributions[iAttribute][iParent].getProbability(iValue));
          text.append(' ');
        }
        text.append('\n');
      }
      text.append("</TABLE>\n");
      text.append("</DEFINITION>\n");
    }
    text.append("</NETWORK>\n");
    text.append("</BIF>\n");
    return text.toString();
  } // toXMLBIF03
  
  
  /** XMLNormalize converts the five standard XML entities in a string
   * g.e. the string V&D's is returned as V&amp;D&apos;s
   * @param sStr string to normalize
   * @return normalized string
   */
  protected String XMLNormalize(String sStr) {
    StringBuffer sStr2 = new StringBuffer();
    for (int iStr = 0; iStr < sStr.length(); iStr++) {
      char c = sStr.charAt(iStr);
      switch (c) {
        case '&': sStr2.append("&amp;"); break;
        case '\'': sStr2.append("&apos;"); break;
        case '\"': sStr2.append("&quot;"); break;
        case '<': sStr2.append("&lt;"); break;
        case '>': sStr2.append("&gt;"); break;
        default:
          sStr2.append(c);
      }
    }
    return sStr2.toString();
  } // XMLNormalize
  
  
    /**
     * @return a string to describe the UseADTreeoption.
     */
    public String useADTreeTipText() {
        return "When ADTree (the data structure for increasing speed on counts,"
            + " not to be confused with the classifier under the same name) is used"
            + " learning time goes down typically. However, because ADTrees are memory"
            + " intensive, memory problems may occur. Switching this option off makes"
            + " the structure learning algorithms slower, and run with less memory."
            + " By default, ADTrees are used.";
    }

    /**
     * @return a string to describe the SearchAlgorithm.
     */
    public String searchAlgorithmTipText() {
        return "Select method used for searching network structures.";
    }

    /**
     * This will return a string describing the BayesNetEstimator.
     * @return The string.
     */
    public String estimatorTipText() {
        return "Select Estimator algorithm for finding the conditional probability tables"
            + " of the Bayes Network.";
    }

    /**
     * @return a string to describe the BIFFile.
     */
    public String BIFFileTipText() {
        return "Set the name of a file in BIF XML format. A Bayes network learned"
            + " from data can be compared with the Bayes network represented by the BIF file."
            + " Statistics calculated are o.a. the number of missing and extra arcs.";
    }

    /**
     * This will return a string describing the classifier.
     * @return The string.
     */
    public String globalInfo() {
        return 
            "Bayes Network learning using various search algorithms and "
          + "quality measures.\n"
          + "Base class for a Bayes Network classifier. Provides "
          + "datastructures (network structure, conditional probability "
          + "distributions, etc.) and facilities common to Bayes Network "
          + "learning algorithms like K2 and B.\n\n"
          + "For more information see:\n\n"
          + "http://www.cs.waikato.ac.nz/~remco/weka.pdf";
    }

    /**
     * Main method for testing this class.
     * 
     * @param argv the options
     */
    public static void main(String[] argv) {
        runClassifier(new BayesNet(), argv);
    } // main

    /** get name of the Bayes network
     * @return name of the Bayes net
     */
    public String getName() {
        return m_Instances.relationName();
    }

    /** get number of nodes in the Bayes network
     * @return number of nodes
     */
    public int getNrOfNodes() {
        return m_Instances.numAttributes();
    }

    /** get name of a node in the Bayes network
     * @param iNode index of the node
     * @return name of the specified node
     */
    public String getNodeName(int iNode) {
        return m_Instances.attribute(iNode).name();
    }

    /** get number of values a node can take
     * @param iNode index of the node
     * @return cardinality of the specified node
     */
    public int getCardinality(int iNode) {
        return m_Instances.attribute(iNode).numValues();
    }

    /** get name of a particular value of a node
     * @param iNode index of the node
     * @param iValue index of the value
     * @return cardinality of the specified node
     */
    public String getNodeValue(int iNode, int iValue) {
        return m_Instances.attribute(iNode).value(iValue);
    }

    /** get number of parents of a node in the network structure
     * @param iNode index of the node
     * @return number of parents of the specified node
     */
    public int getNrOfParents(int iNode) {
        return m_ParentSets[iNode].getNrOfParents();
    }

    /** get node index of a parent of a node in the network structure
     * @param iNode index of the node
     * @param iParent index of the parents, e.g., 0 is the first parent, 1 the second parent, etc.
     * @return node index of the iParent's parent of the specified node
     */
    public int getParent(int iNode, int iParent) {
        return m_ParentSets[iNode].getParent(iParent);
    }

	/** Get full set of parent sets.
	 * @return parent sets;
	 */
	public ParentSet[] getParentSets() { 
		return m_ParentSets;
	}

	/** Get full set of estimators.
	 * @return estimators;
	 */
	public Estimator[][] getDistributions() {
		return m_Distributions;
	}

    /** get number of values the collection of parents of a node can take
     * @param iNode index of the node
     * @return cardinality of the parent set of the specified node
     */
    public int getParentCardinality(int iNode) {
        return m_ParentSets[iNode].getCardinalityOfParents();
    }

    /** get particular probability of the conditional probability distribtion
     * of a node given its parents.
     * @param iNode index of the node
     * @param iParent index of the parent set, 0 <= iParent <= getParentCardinality(iNode)
     * @param iValue index of the value, 0 <= iValue <= getCardinality(iNode)
     * @return probability
     */
    public double getProbability(int iNode, int iParent, int iValue) {
        return m_Distributions[iNode][iParent].getProbability(iValue);
    }

    /** get the parent set of a node 
     * @param iNode index of the node
     * @return Parent set of the specified node.
     */
    public ParentSet getParentSet(int iNode) {
        return m_ParentSets[iNode];
    }

	/** get ADTree strucrture containing efficient representation of counts.
	 * @return ADTree strucrture
	 */
	public ADNode getADTree() { return m_ADTree;}

	// implementation of AdditionalMeasureProducer interface
	  /**
	   * Returns an enumeration of the measure names. Additional measures
	   * must follow the naming convention of starting with "measure", eg.
	   * double measureBlah()
	   * @return an enumeration of the measure names
	   */
	  public Enumeration enumerateMeasures() {
	    Vector newVector = new Vector(4);
	    newVector.addElement("measureExtraArcs");
	    newVector.addElement("measureMissingArcs");
	    newVector.addElement("measureReversedArcs");
	    newVector.addElement("measureDivergence");
	    newVector.addElement("measureBayesScore");
	    newVector.addElement("measureBDeuScore");
	    newVector.addElement("measureMDLScore");
	    newVector.addElement("measureAICScore");
	    newVector.addElement("measureEntropyScore");
	    return newVector.elements();
	  } // enumerateMeasures

	  public double measureExtraArcs() {
	  	if (m_otherBayesNet != null) {
	  		return m_otherBayesNet.extraArcs(this); 
	  	}
	  	return 0;
	  } // measureExtraArcs

	  public double measureMissingArcs() {
	  	if (m_otherBayesNet != null) {
	  		return m_otherBayesNet.missingArcs(this); 
	  	}
	  	return 0;
	  } // measureMissingArcs

  	  public double measureReversedArcs() {
	  	if (m_otherBayesNet != null) {
	  		return m_otherBayesNet.reversedArcs(this); 
	  	}
	  	return 0;
	  } // measureReversedArcs
	  	  
  	  public double measureDivergence() {
	  	if (m_otherBayesNet != null) {
	  		return m_otherBayesNet.divergence(this); 
	  	}
	  	return 0;
	  } // measureDivergence

  	  public double measureBayesScore() {
  	  	LocalScoreSearchAlgorithm s = new LocalScoreSearchAlgorithm(this, m_Instances);
  	  	return s.logScore(Scoreable.BAYES);
  	  } // measureBayesScore

  	  public double measureBDeuScore() {
  	  	LocalScoreSearchAlgorithm s = new LocalScoreSearchAlgorithm(this, m_Instances);
  	  	return s.logScore(Scoreable.BDeu);
  	  } // measureBDeuScore

  	  public double measureMDLScore() {
  	  	LocalScoreSearchAlgorithm s = new LocalScoreSearchAlgorithm(this, m_Instances);
  	  	return s.logScore(Scoreable.MDL);
  	  } // measureMDLScore

  	  public double measureAICScore() {
  	  	LocalScoreSearchAlgorithm s = new LocalScoreSearchAlgorithm(this, m_Instances);
  	  	return s.logScore(Scoreable.AIC);
  	  } // measureAICScore

  	  public double measureEntropyScore() {
  	  	LocalScoreSearchAlgorithm s = new LocalScoreSearchAlgorithm(this, m_Instances);
  	  	return s.logScore(Scoreable.ENTROPY);
  	  } // measureEntropyScore
  	  
  	  /**
	   * Returns the value of the named measure
	   * @param measureName the name of the measure to query for its value
	   * @return the value of the named measure
	   * @throws IllegalArgumentException if the named measure is not supported
	   */
	  public double getMeasure(String measureName) {
	  	if (measureName.equals("measureExtraArcs")) {
	  		return measureExtraArcs();
	  	}
	  	if (measureName.equals("measureMissingArcs")) {
	  		return measureMissingArcs();
	  	}
	  	if (measureName.equals("measureReversedArcs")) {
	  		return measureReversedArcs();
	  	}
	  	if (measureName.equals("measureDivergence")) {
	  		return measureDivergence();
	  	}
	  	if (measureName.equals("measureBayesScore")) {
	  		return measureBayesScore();
	  	}
	  	if (measureName.equals("measureBDeuScore")) {
	  		return measureBDeuScore();
	  	}
	  	if (measureName.equals("measureMDLScore")) {
	  		return measureMDLScore();
	  	}
	  	if (measureName.equals("measureAICScore")) {
	  		return measureAICScore();
	  	}
	  	if (measureName.equals("measureEntropyScore")) {
	  		return measureEntropyScore();
	  	}
	  	return 0;
	  } // getMeasure

} // class BayesNet