edu.stanford.nlp.parser.metrics
Class DependencyEval
java.lang.Object
edu.stanford.nlp.parser.metrics.AbstractEval
edu.stanford.nlp.parser.metrics.DependencyEval
public class DependencyEval
- extends AbstractEval
Evaluates the dependency accuracy of a tree (based on HeadFinder
dependency judgments).
TODO CDM Mar 2004: This should be rewritten so as to root a word at an
index position; otherwise it doesn't work correctly when you get two
identical dependents (like with "I went to Greece to see the ruins").
TODO spence Feb 2010: Trees not only need to implement the HasIndex
interface, but also the head finding algorithm must set the index()
annotation for each phrasal node according to the head.
- Author:
- Dan Klein
Method Summary |
protected Set<?> |
makeObjects(Tree tree)
Build the set of dependencies for evaluation. |
Methods inherited from class edu.stanford.nlp.parser.metrics.AbstractEval |
display, display, evaluate, evaluate, evaluate, getEvalbF1, getEvalbF1Percent, getExact, getExactPercent, getLastF1, getNum, getSentAveF1, precision |
Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
DependencyEval
public DependencyEval(String str,
boolean runningAverages,
Filter<String> punctFilter)
- Parameters:
punctFilter
- A filter that accepts punctuation words.
makeObjects
protected Set<?> makeObjects(Tree tree)
- Build the set of dependencies for evaluation. This set excludes
all dependencies for which the argument is a punctuation tag.
- Specified by:
makeObjects
in class AbstractEval
Stanford NLP Group