単語の用例の半教師有りクラスタリング  [in Japanese] Semi-Supervised Clustering For Word Examples  [in Japanese]

Access this Article

Search this Article

Author(s)

Abstract

単語の用例をクラスタリングすることは,教師有りの語義曖昧性解消手法のためのタグ付きコーパス作成支援,新語義候補の抽出,多義性解消システムの精度改善,などに有効であると考えられる.本研究では,この単語の用例のクラスタリングに,半教師有りクラスタリングを適用する.我々の提案する半教師有りクラスタリングは,種用例間に導入する制約に関して,''cannot-link''の制約を重視していること,また,語義タグを付与した種用例を含むクラスタの重心の変動を抑えること,において新規性がある.本論文では,この提案手法を「SENSEVAL-2 日本語辞書タスク」のデータに適用した結果について報告する.Clustering for examples of a word is effective in supporting to construct tagged corpus for supervised word sense disambiguation, extracting candidates for a new word sense, improving accuracy in a word sense disambiguation system. In our study, we apply semi-supervised clustering approach to cluster examples of a word. Our proposed semi-supervised clustering approach is novel in that we focus on ''cannot-link'' with regard to constraints between seed examples and control the fluctuation of the centroid of a cluster. In this paper, we report the results obtained by applying our proposed method to the data of ''SENSEVAL-2 Japanese dictionary task.''

Clustering for examples of a word is effective in supporting to construct tagged corpus for supervised word sense disambiguation, extracting candidates for a new word sense, improving accuracy in a word sense disambiguation system. In our study, we apply semi-supervised clustering approach to cluster examples of a word. Our proposed semi-supervised clustering approach is novel in that we focus on "cannot-link" with regard to constraints between seed examples and control the fluctuation of the centroid of a cluster. In this paper, we report the results obtained by applying our proposed method to the data of "SENSEVAL-2 Japanese dictionary task."

Journal

  • IPSJ SIG Notes

    IPSJ SIG Notes 2008(33(2008-NL-184)), 7-12, 2008-03-27

    Information Processing Society of Japan (IPSJ)

References:  14

Codes

  • NII Article ID (NAID)
    110006825037
  • NII NACSIS-CAT ID (NCID)
    AN10115061
  • Text Lang
    JPN
  • Article Type
    Technical Report
  • ISSN
    09196072
  • NDL Article ID
    9456928
  • NDL Source Classification
    ZM13(科学技術--科学技術一般--データ処理・計算機)
  • NDL Call No.
    Z14-1121
  • Data Source
    CJP  NDL  NII-ELS  IPSJ 
Page Top