統計的アプローチによる英語スラッシュ・リーディング教材の自動生成(自然言語)  [in Japanese] Automatic Generation of Materials for Slash Reading Based on Statistical Models(Natural Language Processing)  [in Japanese]

Abstract

スラッシュ・リーディングとは,意味のかたまりごとにスラッシュで区切られた英文を読むことにより,読解力の向上を目指す学習法である.多くのスラッシュ付き英文を読むことで,学習効果が上がると考えられるが,現在のところ十分な文書数のある学習教材が存在しないという問題がある.本稿では,統計的アプローチを用いて任意の英文にスラッシュを自動的に挿入する手法を提案する.英文中のスラッシュの位置を定める主な要因は,英文の部分的な構文構造・セグメント長のバランス・一部の単語であるという仮定に基づき,パラメトリックな確率モデルおよびSVMを構築する.既存の教材を学習データとしてモデルを学習することで,その教材のスラッシュ挿入規則を模倣したスラッシュ付き英文を作ることができる.3つの既存教材を対象とした実験では,提案手法が,様々な教材におけるスラッシュ挿入規則を,従来手法よりも高い適合率・再現率で模倣できるという結果が示されている.

In Slash Reading, learners read English sentences separated into segments (sense groups) with slashes to improve their reading skills. The more texts for Slash Reading a learner read, the more effect of learning could be expected. However, there are not enough materials for Slash Reading. This paper proposes methods for transforming automatically a plain sentence into a slashed sentence based on statistical approaches. A parametric model and a SVM model are built on the assumption that the factors to decide where to insert slashes into a sentence are a portion of the syntactic structure of the sentence, the lengths of the segments and words around the slashes. The models are learned from an existing material for Slash Reading. The systems based on these models, therefore, can transform automatically a plain sentence into a slashed sentence by imitating positions of slashes in the material. The results of the experiments using existing materials for Slash Reading indicate that the proposed methods imitate positions of slashes of the materials with the higher precision and recall than the previous methods.

Journal

IPSJ Journal   [List of Volumes]

IPSJ Journal 48(1), 365-374, 2007-01-15  [Table of Contents]

Information Processing Society of Japan (IPSJ)

References:  9

You must have a user ID to see the references.If you already have a user ID, please click "Login" to access the info.New users can click "Sign Up" to register for an user ID.

Cited by:  1

You must have a user ID to see the cited references.If you already have a user ID, please click "Login" to access the info.New users can click "Sign Up" to register for an user ID.

Preview

Preview

Codes

  • NII Article ID (NAID) :
    110006152210
  • NII NACSIS-CAT ID (NCID) :
    AN00116647
  • Text Lang :
    JPN
  • Article Type :
    Journal Article
  • ISSN :
    03875806
  • NDL Article ID :
    8649992
  • 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. :
    Z14-741
  • Databases :
    CJP  CJPref  NDL  NII-ELS 

Share