uk.ac.essex.malexa.nlp.dp.GuiTAR.eval
Class AnaphoraResolutionEvaluation

java.lang.Object
  extended byuk.ac.essex.malexa.nlp.dp.GuiTAR.eval.AnaphoraResolutionEvaluation

public class AnaphoraResolutionEvaluation
extends Object

A class that encapsulates the functionality for evaluating the performance of the Anaphora Resolution Engine running over an annotated corpus.

Version:
1.0
Author:
Mijail A. Kabadjov

Constructor Summary
AnaphoraResolutionEvaluation()
           
 
Method Summary
private  String getInformationString(Element element)
          A method that given a DOM node of an NP, extracts 6 features: 1.
 void processFile(String corpusFileName, String inputFileName, Vector categoryName)
          Given a corpus file and an automatically parsed and annotated file (1.txtToXML, 2.ARE_Runner), this method, by first aligning the corpus nes with the parsed ones, computes and displays performance statistics for the automatic annotation in terms of the parsed nes with reference to the corpus annotation.
 void processFile(String inputFileName, Vector categoryName)
          Given a corpus file, which contains original annotation and automatic annotation, this method computes and displays performance statistics of the latter with reference to the former.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

AnaphoraResolutionEvaluation

public AnaphoraResolutionEvaluation()
Method Detail

processFile

public void processFile(String inputFileName,
                        Vector categoryName)
Given a corpus file, which contains original annotation and automatic annotation, this method computes and displays performance statistics of the latter with reference to the former.

Parameters:
inputFileName - The name of the file to be processed
categoryName - A vector of types of NPs to be considered (pers-pro, the-np, etc.)

getInformationString

private String getInformationString(Element element)
A method that given a DOM node of an NP, extracts 6 features: 1. NP string, 2. NP type (original), 3. NP type (by heuristics), 4. Head, 5. Modifiers, 6. Utterance id

Parameters:
element - The DOM node holgind the NP to be processed
Returns:
String The list of extracted features separated by "^"

processFile

public void processFile(String corpusFileName,
                        String inputFileName,
                        Vector categoryName)
Given a corpus file and an automatically parsed and annotated file (1.txtToXML, 2.ARE_Runner), this method, by first aligning the corpus nes with the parsed ones, computes and displays performance statistics for the automatic annotation in terms of the parsed nes with reference to the corpus annotation.

Parameters:
corpusFileName - The name of the file from which corpus annotation will be extracted
inputFileName - The name of the automatically parsed and annotated file
categoryName - A vector of types of NPs to be considered (pers-pro, the-np, etc.)