Web からの自己教師あり学習を用いた人間行動マイニング  [in Japanese] Self-Supervised Mining Human Activity from the Web  [in Japanese]

Search this Article

Author(s)

Abstract

本論文の目的は,日本語Webページの文中に現れる行動の基本属性(行動主,動作,対象,時刻,場面,場所)と行動間の遷移を自動的に抽出することである.しかし,先行研究では,抽出のための準備コストが大きいことや,抽出できる行動属性が少ないこと,適用可能な文の種類が少ないこと,行動属性間の係り受け関係を十分に考慮されていないこと,そしてプライバシーなどといった問題がある.そこで本論文では,条件付確率場(Conditional Random Fields)と自己教師あり学習(Self-Supervised Learning)を用いて,行動属性と行動間の遷移を自動的に抽出する手法を提案する.提案手法では,人手でラベル編集,初期インスタンスの作成,行動のドメインの定義などの必要がなく,一回のテストで文中に現れる行動属性と行動間の遷移を漏れなく全て抽出でき,高い精度が得られる(行動:88.9%,基本行動属性:90%以上,行動間の遷移:87.5%).

In our definition, human activity can be expressed by five basic attributes : actor, action, object, time and location. The goal of this paper is describe a novel method to automatically extract all of the basic attributes and the transition between activities derived from sentences in Japanese web pages. However, previous work had some limitations, such as high setup costs, inability to extract all attributes, limitation on the types of sentences that can be handled, and insufficient consideration interdependency among attributes. To resolve these problems, this paper proposes a novel approach that uses conditional random fields and self-supervised learning. This approach treats activity extraction as a sequence labeling problem, and has advantages such as domain-independence, scalability, and does not require any human input. Since it is unnecessary to fix the number of elements in a tuple, this approach can extract all of the basic attributes and the transition between activities by making only a simgle pass. In an experiment, this approach achieves high precision (activity : 88.9%, attributes : over 90%, transition : over 87%).

Journal

  • IEICE technical report

    IEICE technical report 109(386), 19-24, 2010-01-15

    The Institute of Electronics, Information and Communication Engineers

References:  20

Cited by:  1

Codes

  • NII Article ID (NAID)
    110008000228
  • NII NACSIS-CAT ID (NCID)
    AN10013061
  • Text Lang
    JPN
  • Article Type
    Journal Article
  • ISSN
    09135685
  • NDL Article ID
    10555010
  • NDL Source Classification
    ZN33(科学技術--電気工学・電気機械工業--電子工学・電気通信)
  • NDL Call No.
    Z16-940
  • Data Source
    CJP  CJPref  NDL  NII-ELS 
Page Top