Archives

期刊全文

Content Analysis - A Corpus Based Model

文件內容之分析--語料庫為本的模型

Volume / Date

 VOL. 11 / 1996

Page

 pp. 95-112

Language

 Chinese

Kuang-Hua Chen 陳光華、Hsin-Hsi Chen 陳信希

Kuang-Hua Chen 陳光華 Department of Library and Information Science, National Taiwan University
國立臺灣大學圖書資訊學系
ORCID: 0000-0003-0616-3815
khchen@ccms.ntu.edu.tw
Hsin-Hsi Chen 陳信希 Department of Computer Science and Information Engineering, National Taiwan University
國立臺灣大學資訊工程學系
ORCID: 0000-0001-9757-9423

Abstract

An important step to understand text is to build the discourse structure through cohesion and coherence. However, to build the discourse structure in turn depends on the full understanding of texts, so that many efforts on this line are not automatic and not successful. A corpus-based model based on 1) repetition of words, 2) importance of words, and 3) collocational semantics for texts is proposed in this paper. It focuses on association norms of noun-noun relations and noun-verb relations defined on discourse level and sentence level, respectively. According to this model, a text partition algorithm is proposed to determine the boundaries of discourse structures and a topic identification algorithm is also presented. The results of a series of experiments show that the proposed model is promising.


一般資訊檢索的研究著重於檢索模型的建構、查詢的回饋機制、檢索行為的探討，檢索系統的執行效能。本文則把研究的重心回歸資訊或文件本身，希望對資訊的內容有一個初步的瞭解。本文根據三個因素：1)詞集的重複，2)詞集的重要性，3)共容語意，提出一個基於真實語料的文件內容分析的模型。這樣的模型著重於文章中名詞／動詞與名詞／名詞之間的配對關係。本文也說明如何使用文件分析模型進行文件切分與文件主題辨識的研究，同時討論相關實驗的結果。

Keywords

Discourse Analysis; Information Retrieval; Natural Language Processing

言談分析、資訊檢索、自然語言處理