bifreader.java

MacroWeka扩展了著名数据挖掘工具weka

                    } catch (NumberFormatException e) {
                    	System.err.println("Wrong number format in position :(" + sX + "," + sY +")");
                   	    m_nPositionX[iNode] = 0;
                   	    m_nPositionY[iNode] = 0;
                    }
                }
            }

        }
        
 		m_Instances = new Instances(sName, attInfo, 100);
		m_Instances.setClassIndex(nNodes - 1);
		setUseADTree(false);
		initStructure();
	} // buildInstances

//	/** selectNodeList selects list of nodes from document specified in XPath expression
//	 * @param doc : document (or node) to query
//	 * @param sXPath : XPath expression
//	 * @return list of nodes conforming to XPath expression in doc
//	 * @throws Exception
//	 */
//	private NodeList selectNodeList(Node doc, String sXPath) throws Exception {
//		NodeList nodelist = org.apache.xpath.XPathAPI.selectNodeList(doc, sXPath);
//		return nodelist;
//	} // selectNodeList

	NodeList selectAllNames(Document doc) throws Exception {
		//NodeList nodelist = selectNodeList(doc, "//NAME");
		NodeList nodelist = doc.getElementsByTagName("NAME");
		return nodelist;
	} // selectAllNames

	NodeList selectAllVariables(Document doc) throws Exception {
		//NodeList nodelist = selectNodeList(doc, "//VARIABLE");
		NodeList nodelist = doc.getElementsByTagName("VARIABLE");
		return nodelist;
	} // selectAllVariables

	Element getDefinition(Document doc, String sName) throws Exception {
		//NodeList nodelist = selectNodeList(doc, "//DEFINITION[normalize-space(FOR/text())=\"" + sName + "\"]");

		NodeList nodelist = doc.getElementsByTagName("DEFINITION");
		for (int iNode = 0; iNode < nodelist.getLength(); iNode++) {
			Node node = nodelist.item(iNode);
			FastVector list = selectElements(node, "FOR");
			if (list.size() > 0) {
				Node forNode = (Node) list.elementAt(0);
				if (getContent((Element) forNode).trim().equals(sName)) {
					return (Element) node;
				}
			}
		}
		throw new Exception("Could not find definition for ((" + sName + "))");
	} // getDefinition

	FastVector getParentNodes(Node definition) throws Exception {
		//NodeList nodelist = selectNodeList(definition, "GIVEN");
		FastVector nodelist = selectElements(definition, "GIVEN");
		return nodelist;
	} // getParentNodes

	String getTable(Node definition) throws Exception {
		//NodeList nodelist = selectNodeList(definition, "TABLE/text()");
		FastVector nodelist = selectElements(definition, "TABLE");
		String sTable = getContent((Element) nodelist.elementAt(0));
		sTable = sTable.replaceAll("\\n"," ");
		return sTable;
	} // getTable

	FastVector selectOutCome(Node item) throws Exception {
		//NodeList nodelist = selectNodeList(item, "OUTCOME");
		FastVector nodelist = selectElements(item, "OUTCOME");
		return nodelist;
	} // selectOutCome

	FastVector selectName(Node item) throws Exception {
	   //NodeList nodelist = selectNodeList(item, "NAME");
	   FastVector nodelist = selectElements(item, "NAME");
	   return nodelist;
   } // selectName

   FastVector selectProperty(Node item) throws Exception {
	  // NodeList nodelist = selectNodeList(item, "PROPERTY");
	  FastVector nodelist = selectElements(item, "PROPERTY");
	  return nodelist;
   } // selectProperty

	FastVector selectElements(Node item, String sElement) throws Exception {
	  NodeList children = item.getChildNodes();
	  FastVector nodelist = new FastVector();
	  for (int iNode = 0; iNode < children.getLength(); iNode++) {
		Node node = children.item(iNode);
		if ((node.getNodeType() == Node.ELEMENT_NODE) && node.getNodeName().equals(sElement)) {
			nodelist.addElement(node);
		}
	  }
	  return nodelist;
  } // selectElements
	/** Count nr of arcs missing from other network compared to current network
	 * Note that an arc is not 'missing' if it is reversed.
	 * @param other: network to compare with
	 * @return nr of missing arcs
	 */
	public int missingArcs(BayesNet other) {
		try {
			Sync(other);
			int nMissing = 0;
			for (int iAttribute = 0; iAttribute < m_Instances.numAttributes(); iAttribute++) {
				for (int iParent = 0; iParent < m_ParentSets[iAttribute].getNrOfParents(); iParent++) {
					int nParent = m_ParentSets[iAttribute].getParent(iParent);
					if (!other.getParentSet(m_order[iAttribute]).contains(m_order[nParent]) && !other.getParentSet(m_order[nParent]).contains(m_order[iAttribute])) {
						nMissing++;
					}
				}
			}
			return nMissing;
		} catch (Exception e) {
			System.err.println(e.getMessage());
			return 0;
		}
	} // missingArcs

	/** Count nr of exta arcs  from other network compared to current network
	 * Note that an arc is not 'extra' if it is reversed.
	 * @param other: network to compare with
	 * @return nr of missing arcs
	 */
	public int extraArcs(BayesNet other) {
		try {
			Sync(other);
			int nExtra = 0;
			for (int iAttribute = 0; iAttribute < m_Instances.numAttributes(); iAttribute++) {
				for (int iParent = 0; iParent < other.getParentSet(m_order[iAttribute]).getNrOfParents(); iParent++) {
					int nParent = m_order[other.getParentSet(m_order[iAttribute]).getParent(iParent)];
					if (!m_ParentSets[iAttribute].contains(nParent) && !m_ParentSets[nParent].contains(iAttribute)) {
						nExtra++;
					}
				}
			}
			return nExtra;
		} catch (Exception e) {
			System.err.println(e.getMessage());
			return 0;
		}
	} // extraArcs


	/** calculates the divergence between the probability distribution
	 * represented by this network and that of another, that is,
	 * \sum_{x\in X} P(x)log P(x)/Q(x)
	 * where X is the set of values the nodes in the network can take,
	 * P(x) the probability of this network for configuration x
	 * Q(x) the probability of the other network for configuration x
	 * @param other: network to compare with
	 * @return divergence between this and other Bayes Network
	 */
	public double divergence(BayesNet other) {
		try {
			Sync(other);
			// D: divergence
			double D = 0.0;
			int nNodes = m_Instances.numAttributes();
			int [] nCard = new int[nNodes];
			for (int iNode = 0; iNode < nNodes; iNode++) {
				nCard[iNode] = m_Instances.attribute(iNode).numValues();
			}
			// x: holds current configuration of nodes
			int [] x = new int[nNodes];
			// simply sum over all configurations to calc divergence D
			int i = 0;
			while (i < nNodes) {
				// update configuration
				x[i]++;
				while (i < nNodes && x[i] == m_Instances.attribute(i).numValues()) {
					x[i] = 0;
					i++;
					if (i < nNodes){
						x[i]++;
					}
				}
				if (i < nNodes) {
					i = 0;
					// calc P(x) and Q(x)
					double P = 1.0;
					for (int iNode = 0; iNode < nNodes; iNode++) {
						int iCPT = 0;
						for (int iParent = 0; iParent < m_ParentSets[iNode].getNrOfParents(); iParent++) {
					    	int nParent = m_ParentSets[iNode].getParent(iParent);
						    iCPT = iCPT * nCard[nParent] + x[nParent];
						} 
						P = P * m_Distributions[iNode][iCPT].getProbability(x[iNode]);
					}
	
					double Q = 1.0;
					for (int iNode = 0; iNode < nNodes; iNode++) {
						int iCPT = 0;
						for (int iParent = 0; iParent < other.getParentSet(m_order[iNode]).getNrOfParents(); iParent++) {
					    	int nParent = m_order[other.getParentSet(m_order[iNode]).getParent(iParent)];
						    iCPT = iCPT * nCard[nParent] + x[nParent];
						} 
						Q = Q * other.m_Distributions[m_order[iNode]][iCPT].getProbability(x[iNode]);
					}
	
					// update divergence if probabilities are positive
					if (P > 0.0 && Q > 0.0) {
						D = D + P * Math.log(Q / P);
					}
				}
			}
			return D;
		} catch (Exception e) {
			System.err.println(e.getMessage());
			return 0;
		}
	} // divergence

	/** Count nr of reversed arcs from other network compared to current network
	 * @param other: network to compare with
	 * @return nr of missing arcs
	 */
	public int reversedArcs(BayesNet other) {
		try {
			Sync(other);
			int nReversed = 0;
		    for (int iAttribute = 0; iAttribute < m_Instances.numAttributes(); iAttribute++) {
				for (int iParent = 0; iParent < m_ParentSets[iAttribute].getNrOfParents(); iParent++) {
					int nParent = m_ParentSets[iAttribute].getParent(iParent);
					if (!other.getParentSet(m_order[iAttribute]).contains(m_order[nParent]) && other.getParentSet(m_order[nParent]).contains(m_order[iAttribute])) {
						nReversed++;
					}
				}
			}
			return nReversed;
		} catch (Exception e) {
			System.err.println(e.getMessage());
			return 0;
		}
	} // reversedArcs

	public BIFReader() {}
		
    public static void main(String[] args) {
        try {
            BIFReader br = new BIFReader();
            br.processFile(args[0]);
	    System.out.println(br.toString());
        
        }
        catch (Throwable t) {
            t.printStackTrace();
        }
    } // main
} // class BIFReader