公开(公告)号：JP2005322161A

公开(公告)日：2005-11-17

申请号：JP2004141515

申请日：2004-05-11

Applicant: Internatl Business Mach Corp ,  インターナショナル・ビジネス・マシーンズ・コーポレーションInternational Business Maschines Corporation

Inventor： SAKUMA ATSUSHI ,  MICHAEL EDWARD FUURU

IPC: G06F17/30 ,  G06K9/62

CPC classification number: G06K9/6219

Abstract: PROBLEM TO BE SOLVED: To provide a selecting device, a selecting method, a program and a recording medium which solve the problem that it is difficult to appropriately select only desired data when there is a large amount of electronized data even though various data are electronized and used in accordance with the advancement and popularization of computers and can efficiently obtain a set related strongly to a predetermined reference element among a plurality of elements stored in a database or the like.
SOLUTION: This selecting device 10 for selecting an adjacent element set whose strength of a relation to the predetermined reference element satisfies a predetermined condition among the plurality of elements is provided with an adjacent element set candidate selecting part 110 for selecting a set of elements whose relation to the reference element is determined to be stronger among the plurality of elements as each of a plurality of adjacent element set candidates being a plurality of candidates in which the adjacent element sets are different from one another, and an adjacent element set selecting part 130 for selecting a set of elements whose relational strength to the reference element satisfies the predetermined condition as an adjacent element set in a sum of sets of the plurality of adjacent element set candidates.
COPYRIGHT: (C)2006,JPO&NCIPI

Abstract translation: 要解决的问题：提供一种选择装置，选择方法，程序和记录介质，其解决了当存在大量的电子化数据时难以适当地仅选择期望数据的问题，即使各种 根据计算机的推进和普及，数据被电子化和使用，并且可以有效地获得与存储在数据库等中的多个元素中的预定参考元素强烈相关的集合。 解决方案：用于选择与预定参考元素的关系的强度在多个元素中满足预定条件的相邻元素集合的选择装置10设置有相邻元素集候选选择部分110，用于选择一组 确定与参考元素的关系的元素在多个元素中更强，因为多个相邻元素集候选中的每一个是相邻元素集彼此不同的多个候选，并且相邻元素集合选择 部分130，用于选择与参考元素的关系强度满足预定条件的一组元素作为在多个相邻元素集合候选中的集合的集合中设置的相邻元素。 版权所有（C）2006，JPO＆NCIPI

公开(公告)号：JP2005301786A

公开(公告)日：2005-10-27

申请号：JP2004118758

申请日：2004-04-14

Applicant: Internatl Business Mach Corp ,  インターナショナル・ビジネス・マシーンズ・コーポレーションInternational Business Maschines Corporation

Inventor： MICHAEL EDWARD FUURU

IPC: G06F17/30 ,  G06F19/00 ,  G06K9/62

CPC classification number: G06K9/6218 ,  G06F17/30705

Abstract: PROBLEM TO BE SOLVED: To select, as a cluster, member elements whose mutual association is strong from among a plurality of elements stored in a data base or the like. SOLUTION: An evaluating apparatus, which calculates a self-confidence value of cluster selection with respect to a cluster generated by selecting any of the plurality of elements, comprises: an evaluated object cluster selection part for selecting as a evaluated cluster an adjacent elements set to be a set of the reference numbers of elements whose association with a reference element is stronger; an adjacent element set selecting part for selecting the adjacent element set to be a set of the reference numbers of elements whose association with the member elements is stronger with respect to the respective member elements included in the cluster; and a confidence value calculating part for calculating the number of elements included in common in the adjacent element set of one member element of a pair and the adjacent element set of the other member element with respect to the pair of the two members selectable from the cluster and outputting the average ratio with respect to the combination of all the member elements as the self-confidence value. COPYRIGHT: (C)2006,JPO&NCIPI

Abstract translation: 要解决的问题：从存储在数据库等中的多个元素中选择相互关联强的成员元素作为集群。 解决方案：一种评估装置，其计算相对于通过选择多个元素中的任何一个而生成的集群的集群选择的自信度值，包括：评估对象集群选择部分，用于选择相邻的 元素被设置为与参考元素的关联更强的元素的参考数的集合; 相邻元素集合选择部分，用于选择相邻元素集合作为与元素元素相关联的元素的参考数量的集合相对于包含在该群集中的各个元素元素更强; 以及置信值计算部，用于计算在一组中的一个成员元素的相邻元素组中相同的元素的数量以及相对于从所述簇中可选择的两个成员对中的另一个元素的相邻元素集合中包含的元素的数量 并输出相对于所有成员元素的组合的平均比例作为自信心值。 版权所有（C）2006，JPO＆NCIPI

公开(公告)号：JP2004199472A

公开(公告)日：2004-07-15

申请号：JP2002368276

申请日：2002-12-19

Applicant: Internatl Business Mach Corp ,  インターナショナル・ビジネス・マシーンズ・コーポレーションInternational Business Maschines Corporation

Inventor： MICHAEL EDWARD FUURU

IPC: G06F17/30

CPC classification number: G06F17/30705 ,  G06F17/30011 ,  Y10S707/99933 ,  Y10S707/99943

Abstract: PROBLEM TO BE SOLVED: To provide a system and a method based on the identification of a cluster of elements (e.g. documents) indicating a degree of high mutual similarity with a background for information retrieval and data mining of a large-scale text database. SOLUTION: The computer system comprises a neighborhood patch generation part 34 for generating data structure for the information retrieval of documents stored in the database and generating a node group having prescribed similarities in a hierarchy structure. The neighborhood patch generation part 34 comprises a hierarchy generation part 36 for generating hierarchy structure upon document-keyword vectors and a patch definition part 26. The computer system comprises a cluster estimation part 28 for generating cluster data of the document-keyword vectors by using the similarities of patches. COPYRIGHT: (C)2004,JPO&NCIPI

Abstract translation: 要解决的问题：提供一种系统和方法，该系统和方法基于识别用于大规模文本的信息检索和数据挖掘的背景的高相互相关程度的元素（例如文档）的集合 数据库。 解决方案：计算机系统包括邻近补丁生成部分34，用于生成用于信息检索存储在数据库中的文档的数据结构，并生成具有规定的层次结构相似性的节点组。 邻域补丁生成部34包括用于根据文档关键词向量生成层次结构的层级生成部36和补丁定义部26.计算机系统包括：群估计部28，用于通过使用文档关键词向量来生成文档关键词向量的群集数据 补丁的相似之处 版权所有（C）2004，JPO＆NCIPI

公开(公告)号：JP2005050197A

公开(公告)日：2005-02-24

申请号：JP2003282690

申请日：2003-07-30

Applicant: Internatl Business Mach Corp ,  インターナショナル・ビジネス・マシーンズ・コーポレーションInternational Business Maschines Corporation

Inventor： KOBAYASHI MEI ,  MICHAEL EDWARD FUURU ,  AONO MASAKI

IPC: G06F17/16 ,  G06F7/00 ,  G06F12/00 ,  G06F17/30

CPC classification number: G06F17/30631

Abstract: PROBLEM TO BE SOLVED: To provide a computer-executable method for reducing dimensions, a program for making a computer execute the method, a dimension reduction device, and a search engine using the device.
SOLUTION: The dimension reduction device which reduces the dimension of a numerical matrix using a computer to obtain a dimension-reduced matrix for providing information includes a processing part 32 which creates and stores a dimension-reduced matrix or index data for reducing dimensions, using a random average matrix RAV. The processing part 32 includes a shuffle vector creating part 44 for creating shuffle vectors for use as shuffle information, and a non-normal basic vector creating part 46 for creating a non-normal basic vector out of the numerical elements of data vectors designated by the shuffle vectors and for storing the non-normal basic vector.
COPYRIGHT: (C)2005,JPO&NCIPI

Abstract translation: 要解决的问题：提供一种用于减小尺寸的计算机可执行方法，用于使计算机执行该方法的程序，尺寸减小装置和使用该装置的搜索引擎。 解决方案：使用计算机减小数字矩阵的尺寸以获得用于提供信息的尺寸减小矩阵的尺寸减小装置包括处理部分32，其创建并存储用于减小尺寸的尺寸减小的矩阵或索引数据 ，使用随机平均矩阵RAV。 处理部分32包括用于创建用作洗牌信息的随机播放向量的随机播放向量创建部分44，以及用于从由数字矢量指定的数据矢量的数字矢量创建非正常基本矢量的非正常基本矢量创建部分46 洗牌向量和用于存储非正常基本向量。 版权所有（C）2005，JPO＆NCIPI

公开(公告)号：JP2003256477A

公开(公告)日：2003-09-12

申请号：JP2002037842

申请日：2002-02-15

Applicant: IBM

Inventor： MICHAEL EDWARD FUURU ,  KOBAYASHI MEI ,  AONO MASAKI

IPC: G06F17/30 ,  G06F17/00

Abstract: PROBLEM TO BE SOLVED: To provide a method for retrieving information from a large scale database (containing several millions of data) in real time, while controlling trade-off in precision of a retrieved result and a response time thereof to a user. SOLUTION: This method is applicable, for example, to the database including a content of a document modeled in a clearly defined distance capable of calculating a distance between any two points, wherein the pair of documents near in its distance is more similar than a pair of documents apart each other. The method is combined with similarity ranking and/or another method to enhance scalability in information retrieval, detection, raking and tracking. COPYRIGHT: (C)2003,JPO