新疆大学学报(自然科学版)
新疆大學學報(自然科學版)
신강대학학보(자연과학판)
XINJIANG UNIVERSITY JOURNAL(NATURAL SCIENCE EDITION)
2014年
4期
475-480
,共6页
米热吉古丽·热扎克%阿里甫·库尔班
米熱吉古麗·熱扎剋%阿裏甫·庫爾班
미열길고려·열찰극%아리보·고이반
维吾尔语%框架语义%目标词%框架元素%例句辅助标注
維吾爾語%框架語義%目標詞%框架元素%例句輔助標註
유오이어%광가어의%목표사%광가원소%례구보조표주
Uyghur%frame semantic%target%frame element%annotation software
构建维吾尔语框架语义知识库是自然语言处理的基础性工程,目前大部分工作基于人工。从真实语料库中抽取包含将要描述词元的例句,为例句标注以及这些例句的配价模式进行深入研究,从标注好的维吾尔语例句中提取维吾尔语词元和例句填充到维吾尔语框架库,构建维吾尔语词元库和例句库,实现维吾尔语语义框架网络的自动构造之目的。本文对维吾尔语真实语料库中抽取的例句中标注对象进行初步分析,论述了人机交互式基于阿拉伯字符的UFN例句辅助标注系统的功能模块设计和工作流程。介绍了UFN例句辅助标注系统的例句标注界面和词元标注报告、词条报告的自动生成演示界面。
構建維吾爾語框架語義知識庫是自然語言處理的基礎性工程,目前大部分工作基于人工。從真實語料庫中抽取包含將要描述詞元的例句,為例句標註以及這些例句的配價模式進行深入研究,從標註好的維吾爾語例句中提取維吾爾語詞元和例句填充到維吾爾語框架庫,構建維吾爾語詞元庫和例句庫,實現維吾爾語語義框架網絡的自動構造之目的。本文對維吾爾語真實語料庫中抽取的例句中標註對象進行初步分析,論述瞭人機交互式基于阿拉伯字符的UFN例句輔助標註繫統的功能模塊設計和工作流程。介紹瞭UFN例句輔助標註繫統的例句標註界麵和詞元標註報告、詞條報告的自動生成縯示界麵。
구건유오이어광가어의지식고시자연어언처리적기출성공정,목전대부분공작기우인공。종진실어료고중추취포함장요묘술사원적례구,위례구표주이급저사례구적배개모식진행심입연구,종표주호적유오이어례구중제취유오이어사원화례구전충도유오이어광가고,구건유오이어사원고화례구고,실현유오이어어의광가망락적자동구조지목적。본문대유오이어진실어료고중추취적례구중표주대상진행초보분석,논술료인궤교호식기우아랍백자부적UFN례구보조표주계통적공능모괴설계화공작류정。개소료UFN례구보조표주계통적례구표주계면화사원표주보고、사조보고적자동생성연시계면。
Construct Uyghur frame semantic knowledge base is a basic pro ject in natural language pro-cessing, at present most of work is based on artificial. Extract sample sentence from real corpus, which contains the describe lexical unit, for depth studied in annotation sentence and valency patterns of these sentence, extracted Uyghur lexical unit from annotated sentence, and sample sentence filled to frame corpus, construct the Uyghur lexicon corpus and sentence corpus, will achieve the Uyghur frame semantic network automatically construct purpose. In this paper makes a preliminary analysis the annotation object of the sentence which extracted from Uyghur real corpus, discuss the design and work process of the man-machine UFN interactive system that based Arabic character. Introduce the sentence annotation window of the UFN man-machine interactive system and lexical unit report, automatically generated demo of the lexical extry report.