首页> 外国专利> Proper noun relationship estimation apparatus, method, and program

Proper noun relationship estimation apparatus, method, and program

机译:专有名词关系估计装置,方法和程序

摘要

PROBLEM TO BE SOLVED: To estimate the relationship between proper nouns with good accuracy about a pair of proper nouns included in a document.SOLUTION: By a proper noun topic acquisition unit 22, according to a proper noun topic model, probability distribution is acquired which shows a relevance degree with a topic with respect to the proper noun. Then, by a proper noun similarity degree calculation unit 23, a proper noun similarity degree between a pair of proper nouns (X, Y) and a pair of proper nouns (A, B) of a teacher case is calculated. By a document topic acquisition unit 26, according to a document topic model, probability distribution is acquired which shows a relevance degree with the topic with respect to the document. By a document similarity degree calculation unit 27, the document similarity degree between an input document and the teacher case document is calculated. By a combination similarity degree calculation unit 28, based on the proper noun similarity degree and the document similarity degree, a combination similarity degree is calculated. By a similar case selection unit 29, a teacher case similar to the pair of the proper nouns (X, Y) is selected. By an output unit 3, the relationship of the pair of the poper nouns (X, Y) is output.
机译:解决的问题:估计关于文档中包括的一对专有名词的专有名词之间的关系的准确性。解决方案:专有名词主题获取单元22根据专有名词主题模型,获取概率分布,其中显示与专有名词相关的话题的相关程度。然后,通过专有名词相似度计算单元23,计算教师案例的一对专有名词(X,Y)和一对专有名词(A,B)之间的专有名词相似度。通过文档主题获取单元26,根据文档主题模型,获取概率分布,该概率分布示出与主题相对于文档的相关度。通过文档相似度计算单元27,计算输入文档与教师案例文档之间的文档相似度。通过组合相似度计算单元28,基于专有名词相似度和文档相似度,计算组合相似度。通过相似案例选择单元29,选择与一对专有名词(X,Y)相似的教师案例。通过输出单元3,输出一对成对名词(X,Y)的关系。

著录项

  • 公开/公告号JP5854975B2

    专利类型

  • 公开/公告日2016-02-09

    原文格式PDF

  • 申请/专利权人 日本電信電話株式会社;

    申请/专利号JP20120259061

  • 发明设计人 平野 徹;牧野 俊朗;

    申请日2012-11-27

  • 分类号G06F17/27;G06F17/30;

  • 国家 JP

  • 入库时间 2022-08-21 14:39:58

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