A Clustering System for Electronic News Articles using Sentense Structure

• KIYOTA Yoji

  Department of Intelligence Science and Technology Graduate School of Informatics, KYOTO University

• KUROHASHI Sadao

  Department of Intelligence Science and Technology Graduate School of Informatics, KYOTO University

• NAKAMURA Jun-ichi

  Department of Intelligence Science and Technology Graduate School of Informatics, KYOTO University

• NAGAO Makoto

  Department of Intelligence Science and Technology Graduate School of Informatics, KYOTO University

Bibliographic Information

Other Title

• 構文情報を利用した電子ニュース記事のクラスタリングシステムの作成と評価

Abstract

We developed a clustering system for electronic news articles using Kohonen's Self-Organization Map. The advantages of this system are:(1)user can get intended articles by referring a unit which seems to contain them, (2)by parsing sentenses and giving larger weights to some terms using postpositional particles of Japanese, this system can arrange articles to fit to user's purpose. For example, when words followed by postpositional particles "wa" "mo" are given larger weights, news articles are arranged by company names.

Journal

• IPSJ SIG Notes

  IPSJ SIG Notes 126 77-84, 1998-07-23

  Information Processing Society of Japan (IPSJ)

Keywords

• self organization map
• clustering
• parsing
• term weighting
• electronic news article
• natural language processing

Details

• CRID: 1570291227270519936
• NII Article ID: 110002934746
• NII Book ID: AN10115061
• Text Lang: ja
• Data Source:
  • CiNii Articles