首页> 外国专利> Word mapping device, word mapping method, program implementing the word mapping method, and storage medium storing the program

Word mapping device, word mapping method, program implementing the word mapping method, and storage medium storing the program

机译:单词映射装置,单词映射方法,实现该单词映射方法的程序以及存储该程序的存储介质

摘要

PPROBLEM TO BE SOLVED: To provide a word mapping technology for processing a great deal of information sources and improving quality of conceptual vector of estimated different unregistered words. PSOLUTION: A word dividing means 11 applies morpheme analysis to a text set. A conceptual vector estimating means 12 searches a conceptual base 16 which stores words and a set of a pairs of concept vectors of the words to determine whether the word is a registered word is or an unregistered word, forms a text set which include different unregistered word with respect to an arbitrary different unregistered word, acquires a conceptual vector of the different registered word in the set, makes the conceptual vector as a parameter correspond to the different unregistered word, then makes the conceptual vector of the different unregistered word minimizing a sum resulting from adding an average of the conceptual vectors in each of the texts in the set to a square sum of a distance between each conceptual vector all over the texts in the set an estimated conceptual vector of the different unregistered word. PCOPYRIGHT: (C)2008,JPO&INPIT
机译:

要解决的问题:提供一种词映射技术,用于处理大量信息源并提高估计的不同未注册词的概念向量的质量。

解决方案:单词划分装置11将语素分析应用于文本集。概念向量估计装置12搜索存储词的概念库16和该词的概念向量对的集合以确定该词是已注册词是还是未注册词,形成包括不同未注册词的文本集对于任意一个不同的未注册词,获取集合中该不同注册词的概念向量,使该概念向量作为参数对应于该不同未注册词,然后使该不同未注册词的概念向量最小化求和将集合中每个文本中的概念向量的平均值与集合中所有文本中每个概念向量之间的距离的平方和相加,即可得出未注册词的估计概念向量。

版权:(C)2008,日本特许厅&INPIT

著录项

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

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