首页> 外国专利> DATABASE GENERATING DEVICE AND RELATIVE DOCUMENT/ RELATIVE WORD RETRIEVAL DEVICE, DATABASE GENERATING METHOD AND RELATIVE DOCUMENT/RELATIVE WORD RETRIEVING METHOD, AND STORAGE MEDIUM

DATABASE GENERATING DEVICE AND RELATIVE DOCUMENT/ RELATIVE WORD RETRIEVAL DEVICE, DATABASE GENERATING METHOD AND RELATIVE DOCUMENT/RELATIVE WORD RETRIEVING METHOD, AND STORAGE MEDIUM

机译:数据库生成设备和相对文档/相对单词检索设备,数据库生成方法和相对文档/相对单词检索方法以及存储介质

摘要

PROBLEM TO BE SOLVED: To retrieve a relative document and a relative word even in a small memory space.;SOLUTION: A natural language analysis of respective documents included in a document group file is taken and an appearing word, the appearing frequency of the appearing word, and a list of IDs of documents where the word has appeared are extracted and saved in a word list file (S1). Then appearing words in a row direction and document numbers in a column direction are set according to the information of appearing words in the word list file to generate a word-document matrix (S2). The word-document matrix is decomposed by singular values to obtain singular value vectors, a word matrix, and a document matrix (S4), and (k) larger elements are taken out as to the singular vectors, word matrix, and document matrix to generate new singular value vectors, a word matrix, and a document matrix, thereby decreasing the number of dimensions.;COPYRIGHT: (C)2001,JPO
机译:解决的问题:即使在很小的存储空间中也要检索相对文档和相对单词;解决方案:对文档组文件中包含的各个文档进行自然语言分析,然后出现一个单词,出现的频率单词,并将出现单词的文档的ID的列表提取并保存在单词列表文件中(S1)。然后,根据单词列表文件中出现单词的信息设置行方向出现的单词和列方向出现的文档编号,以生成单词文档矩阵(S2)。通过奇异值分解单词文档矩阵以获得奇异值向量,单词矩阵和文档矩阵(S4),并针对奇异矢量,单词矩阵和文档矩阵取出较大的元素(k)生成新的奇异值向量,单词矩阵和文档矩阵,从而减少维数。;版权所有:(C)2001,JPO

著录项

  • 公开/公告号JP2001014341A

    专利类型

  • 公开/公告日2001-01-19

    原文格式PDF

  • 申请/专利权人 RICOH CO LTD;

    申请/专利号JP19990188613

  • 发明设计人 ISHIOKA TSUNENORI;KAMEDA MASAYUKI;

    申请日1999-07-02

  • 分类号G06F17/30;

  • 国家 JP

  • 入库时间 2022-08-22 01:31:35

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