eduzhai > Applied Sciences > Computer Science >

Coherent Keyphrase Extraction via Web Mining

  • paiqiu
  • (0) Download
  • 20210805
  • Save

... pages left unread,continue reading

Document pages: 6 pages

Abstract: Keyphrases are useful for a variety of purposes,including summarizing, indexing, labeling,categorizing, clustering, highlighting, browsing, andsearching. The task of automatic keyphrase extractionis to select keyphrases from within the text of a givendocument. Automatic keyphrase extraction makes itfeasible to generate keyphrases for the huge number ofdocuments that do not have manually assignedkeyphrases. A limitation of previous keyphraseextraction algorithms is that the selected keyphrases areoccasionally incoherent. That is, the majority of theoutput keyphrases may fit together well, but there maybe a minority that appear to be outliers, with no clearsemantic relation to the majority or to each other. Thispaper presents enhancements to the Kea keyphraseextraction algorithm that are designed to increase thecoherence of the extracted keyphrases. The approach isto use the degree of statistical association amongcandidate keyphrases as evidence that they may besemantically related. The statistical association ismeasured using web mining. Experiments demonstratethat the enhancements improve the quality of theextracted keyphrases. Furthermore, the enhancementsare not domain-specific: the algorithm generalizes wellwhen it is trained on one domain (computer sciencedocuments) and tested on another (physics documents).

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×