edu.columbia.cs.cg.prdualrank.pattern.extractor.impl
Class WindowedSearchPatternExtractor<T extends Document>

java.lang.Object
  edu.columbia.cs.cg.prdualrank.pattern.extractor.SearchPatternExtractor<Document>
      edu.columbia.cs.cg.prdualrank.pattern.extractor.impl.WindowedSearchPatternExtractor<T>

All Implemented Interfaces:

PatternExtractor<Document>

public class WindowedSearchPatternExtractor<T extends Document>
extends SearchPatternExtractor<Document>

This class is used for our implementation of: "Searching Patterns for Relation Extraction over the Web: Rediscovering the Pattern-Relation Duality" . Y. Fang and K. C.-C. Chang. In WSDM, pages 825-834, 2011. For further information, WSDM 2011 Conference Website .

Description

This class defines the behavior of the generation of search pattern as discussed with one of the authors of PRDualRank: Yuan Fang.
Notice that it differs from the one described in the paper on the Universe of words used to generate the Search Patterns.
In this case, a window around the tuples is used. However, in the paper, such window is not described.
According to the authors, this approach was used to run the experiments to obtain better performance (in time) of the results. Processing all the words in a document is computationally expensive.
For the original implementation, please see the DocumentSearchPatternExtractor class. Consider that the execution time will increase considerably compared to this implementation.

No definition can be pointed out to the paper due to implementation details described above.

Since:

2011-10-07

Version:

0.1

Author:

Pablo Barrio, Goncalo Simoes

See Also:

WSDM 2011 Conference Website , DocumentSearchPatternExtractor

Constructor Summary

WindowedSearchPatternExtractor(int window, int ngram, int numberOfPhrases)
Instantiates a new windowed search pattern extractor.

Method Summary

Methods inherited from class edu.columbia.cs.cg.prdualrank.pattern.extractor.SearchPatternExtractor

extractPatterns

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

WindowedSearchPatternExtractor

public WindowedSearchPatternExtractor(int window,
                                      int ngram,
                                      int numberOfPhrases)

Instantiates a new windowed search pattern extractor.

Parameters:

window - the number of words around the tuple to be used to generate the search patterns.

ngram - the maximum size of ngrams to be calculated in order to generate the search patterns.

numberOfPhrases - the maximum number of ngrams to be combined in the search pattern generation.