A Study to Improve the Efficiency of a Discourse Parsing System

A Study to Improve the Efficiency of a Discourse Parsing System
复制标题

提高语篇解析系统效率的研究

DOI：
--
复制
发表时间：
2003
期刊：
Conference on Intelligent Text Processing and Computational Linguistics
影响因子：
--
通讯作者：
G. Abeysinghe
G. Abeysinghe
中科院分区：
--
文献类型：
--
作者：
H. T. Le;G. Abeysinghe

文献摘要

被引文献

相似文献

本文研究了一个只考虑重要特征的语篇分析系统的实现。修辞关系的识别基于三种类型的线索短语(普通线索短语、名词短语线索和动词短语线索)和不同的语篇连贯手段。为了建立一个高精度、低复杂度的系统，开发了句法分析算法及其规则集。本系统使用的数据来自语言数据联盟(LDC)的RST话语树库。
This paper presents a study of the implementation of a discourse parsing system, where only significant features are considered. Rhetorical relations are recognized based on three types of cue phrases (the normal cue phrases, Noun-Phrase cues and Verb-Phrase cues), and different textual coherence devices. The parsing algorithm and its rule set are developed in order to create a system with high accuracy and low complexity. The data used in this system are taken from the RST Discourse Treebank of the Linguistic Data Consortium (LDC).