首页> 外文会议>International Conference on Asian Language Processing >Warning and Suggestion System on Syntax Tree Maker Application
【24h】

Warning and Suggestion System on Syntax Tree Maker Application

机译:语法树制造商应用程序的警告和建议系统

获取原文

摘要

TIKSentence TreeMaker is an application built by the Indonesian Association of Computational Linguistics (INACL) which has useful to give phrase labels in sentences. This Application has advantage in User Interface (UI) system, where word's structure in a sentence can be built into tree structure in real time, so users see and do sentence's structure easily. Since every application has advantages and disadvantages, TIKSentence TreeMaker has a disadvantage where this application cannot assist users in giving phrase labels and users can give any labels freely whether it is fit or not to the word's structure. Author try to correct this disadvantage by creating a system that can detect phrase labels and give a recommended phrase label that fir for the word's structure. This system is built to detect phrase labels in syntactic manner with Context Free Grammar (CFG) where it is useful to change application's tree structure phrase into another form that can be readed by computer easily. System check every structure phrase that tested and change its phrase labels if the phrase labels are detected to be incorrect into a phrase label recommended by the system without change its phrase structure. The accuracy result of the built system is around 95% while the initial data test is around 87% which means our system succeds changing some incorrect phrase labels into correct phrase labels.
机译:Tiksentence Triemaker是印度尼西亚计算语言学协会(INACL)构建的应用程序,这有助于在句子中给短语标签。此应用程序在用户界面(UI)系统中具有优势,其中句子中的Word的结构可以实时内置为树结构,因此用户可以轻松地查看和执行句子的结构。由于每个应用程序都有优势和缺点,因此Tiksentence Traemaker具有缺点,其中该应用程序无法帮助用户给出短语标签,用户可以自由地为任何标签提供任何标签,无论是适合这个词的结构。作者尝试通过创建可以检测到短语标签的系统来纠正此缺点,并为单词结构提供推荐的短语标签。构建此系统以通过语法方式检测短语标签,其中包含上下文自由语法(CFG),在那里将应用程序的树结构短语更改为可以轻松读取的另一种形式是有用的。系统检查测试和更改其短语标签的每个结构短语如果检测到该短语标签被检测到系统建议的短语标签而不更改其短语结构。内置系统的准确性结果约为95%,而初始数据测试约为87%,这意味着我们的系统成员将一些不正确的短语标签更改为正确的短语标签。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号