中科微点-全球专利检索

  1. Login
  2. Sign Up

Search Results

14 records (took 0.024 seconds)

    Device, method and program for supporting evaluation work of new customer candidate
    1.
    发明专利
    Device, method and program for supporting evaluation work of new customer candidate 有权
    用于支持新客户候选人评估工作的设备，方法和程序

    公开(公告)号：JP2010152568A

    公开(公告)日：2010-07-08

    申请号：JP2008328913

    申请日：2008-12-25

    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： YANAGISAWA HIROKI ,  KAJIMA HISATSUGU ,  TAJIMA REI

    IPC: G06Q10/00 ,  G06Q30/02 ,  G06Q50/00

    Abstract: PROBLEM TO BE SOLVED: To provide a technology for supporting evaluation work of a new customer candidate. SOLUTION: The device for supporting evaluation work of the new customer candidate is configured to calculate a discrimination surface in an SVM using a group of feature vectors searched for each of a plurality of objects having either a customer label or a non-customer label as training data. Then, the support device is configured to calculate a distance between each of objects listed in a customer candidate list and each of a plurality of objects having either the customer label or the non-customer label as a length obtained by projecting a distance between the feature vectors to the normal vectors of the discrimination surface, and to extract the object positioned in the neighborhood of each object of the customer candidate list and having the customer label according to each calculated distance, and to record the identification information of each extracted object in association with the object of the corresponding customer candidate. COPYRIGHT: (C)2010,JPO&INPIT

    Abstract translation: 要解决的问题：提供支持新客户候选人的评估工作的技术。

    解决方案：用于支持新客户候选者的评估工作的设备被配置为使用搜索到具有客户标签或非客户的多个对象中的每一个的特征向量组来计算SVM中的识别表面 标签作为培训数据。 然后，支持装置被配置为计算在客户候选列表中列出的每个对象之间的距离和具有客户标签或非客户标签的多个对象中的每一个作为通过投影特征之间的距离而获得的长度 向对应于判别面的法线向量的矢量，并且根据每个计算出的距离提取位于客户候选列表的每个对象附近的对象并具有客户标签，并且将每个提取的对象的标识信​​息关联起来 与相应的客户候选人的对象。 版权所有（C）2010，JPO＆INPIT

    Method for predicting work evaluation value, program and system
    2.
    发明专利
    Method for predicting work evaluation value, program and system 有权
    预测工作评估价值，程序和制度的方法

    公开(公告)号：JP2010061323A

    公开(公告)日：2010-03-18

    申请号：JP2008225425

    申请日：2008-09-03

    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： KAJIMA HISATSUGU ,  HITO SHOHEI ,  TAJIMA REI

    IPC: G06Q10/00 ,  G06Q10/06 ,  G06Q50/00

    Abstract: PROBLEM TO BE SOLVED: To provide a method for predicting a work evaluation value of a worker. SOLUTION: At the stage that labeled data of full operations and negligent operations of a plurality of times performed by a plurality of worker is preserved in a computer hard disk, for i=1, ..., M (where M is the number of workers), machine learning based on a linear identification model is applied to labeled data of the full operations and the negligent operations by a worker i. As a result, parameters w (1) , w (2) , ..., w (M) of the respective workers making trial are obtained, and by dividing the sum thereof by M, a mean w is obtained. Alternatively, it may also be possible to obtain a weighted mean w which is weighted by the number of times of trials by each worker. By this, the parameter w of the total results is obtained and this w is applied to an evaluation function. Then, for example, by obtaining an internal product of work data x of a new worker and w, an evaluation value of x is obtained. COPYRIGHT: (C)2010,JPO&INPIT

    Abstract translation: 要解决的问题：提供一种用于预测工作人员的工作评价值的方法。 解决方案：在计算机硬盘中保存多个工作人员执行多次操作的完整操作和疏忽操作的数据的阶段，对于i = 1，...，M（其中M为 工人人数），基于线性识别模型的机器学习应用于工作人员的全面操作和疏忽操作的标签数据。 因此，各试验工作者的参数w （1），w （2），...，w （M） 并且通过将其和除以M，获得平均值w。 或者，也可以获得加权平均值w，该加权平均值w由每个工作者的试验次数加权。 由此得到总结果的参数w，将该w应用于评价函数。 然后，例如，通过获得新工作者的工作数据x的内部积和w，得到x的评价值。 版权所有（C）2010，JPO＆INPIT

    Change analysis system, method, and program
    3.
    发明专利
    Change analysis system, method, and program 有权
    变更分析系统，方法和程序

    公开(公告)号：JP2009205615A

    公开(公告)日：2009-09-10

    申请号：JP2008049729

    申请日：2008-02-29

    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： HITO SHOHEI ,  IDE TAKESHI ,  KAJIMA HISATSUGU ,  KUBO HARUNOBU ,  MATSUZAWA YASUSHI

    IPC: G06F17/30 ,  G06N5/04

    CPC classification number: G06K9/623 ,  G06N99/005

    Abstract: PROBLEM TO BE SOLVED: To provide a method for efficiently solving the change analysis problem.
    SOLUTION: Different virtual labels, for example, like +1 and -1, are assigned to two data sets. A change analysis problem for the two data sets is reduced to a supervised learning problem by using the virtual labels. Specifically, a classifier such as logical regression, decision tree and SVM is prepared and is trained by use of a data set obtained by merging the two data sets assigned the virtual labels. A feature selection function of the resultant classifier is used to rank and output both every attribute contributing to classification and its contribution rate.
    COPYRIGHT: (C)2009,JPO&INPIT

    Abstract translation: 要解决的问题：提供一种有效解决变化分析问题的方法。 解决方案：将不同的虚拟标签（例如+1和-1）分配给两个数据集。 通过使用虚拟标签将两个数据集的变化分析问题简化为监督学习问题。 具体地说，准备了诸如逻辑回归，决策树和SVM的分类器，并且通过使用通过合并分配了虚拟标签的两个数据集获得的数据集进行训练。 使用得到的分类器的特征选择功能对贡献于分类的每个属性及其贡献率进行排序和输出。 版权所有（C）2009，JPO＆INPIT

    Device, method and program for detecting classification factor and recording medium
    4.
    发明专利
    Device, method and program for detecting classification factor and recording medium 有权
    用于检测分类因子和记录介质的设备，方法和程序

    公开(公告)号：JP2005044163A

    公开(公告)日：2005-02-17

    申请号：JP2003278138

    申请日：2003-07-23

    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： INOGUCHI AKIHIRO ,  KAJIMA HISATSUGU

    IPC: C40B30/02 ,  G06F17/00 ,  G06F17/30 ,  G06F19/00

    CPC classification number: G06F19/707 ,  Y10S707/99936 ,  Y10S707/99943

    Abstract: PROBLEM TO BE SOLVED: To detect a set of appropriate conditions for classifying data. SOLUTION: A classification factor detection device detects a set of component elements of objects serving as classification factors as to the results of classifying a plurality of objects. The device includes a first selection means for selecting a first pattern as a set of component elements from a plurality of component elements; a second selection means for selecting a second pattern obtained by the addition of other component elements to the first pattern; an evaluation value creation means by which an evaluation value of such accuracy as to be capable of classifying the plurality of objects based on a classification condition requiring the inclusion of the first pattern and no second pattern is created based on the number of objects meeting the condition, among objects classified as a first group, and the number of objects meeting the condition, among objects classified as a second group; and a classification factor output means which outputs the first and second patterns as classification factors if the accuracy indicated by the evaluation value exceeds reference accuracy. COPYRIGHT: (C)2005,JPO&NCIPI

    Abstract translation: 要解决的问题：检测用于对数据进行分类的一组适当条件。 解决方案：分类因子检测装置检测作为用于对多个对象进行分类的结果的分类因子的对象的一组组成元素。 该装置包括：第一选择装置，用于从多个组成元素中选择第一图案作为一组组成元素; 第二选择装置，用于选择通过向第一图案添加其它组成元素而获得的第二图案; 评估值创建装置，通过该评估值创建装置，基于满足条件的对象的数量，创建能够基于需要包含第一模式的分类条件而不是第二模式来分类多个对象的精度的评估值 在分类为第一组的对象中，在分类为第二组的对象中，满足该条件的对象的数量; 以及如果由评估值指示的精度超过参考精度，则输出第一和第二图案作为分类因子的分类系数输出装置。 版权所有（C）2005，JPO＆NCIPI

    Data processing method, information processing system using it, and program
    5.
    发明专利
    Data processing method, information processing system using it, and program 有权
    数据处理方法，使用它的信息处理系统和程序

    公开(公告)号：JP2003271599A

    公开(公告)日：2003-09-26

    申请号：JP2002064452

    申请日：2002-03-08

    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

    Inventor： KAJIMA HISATSUGU ,  KOYANAGI MITSUO

    IPC: G06F17/10 ,  G06F15/00 ,  G06F17/30

    CPC classification number: G06K9/6282

    Abstract: PROBLEM TO BE SOLVED: To realize the classification of semi-structured data by kernel method to a labelled order tree.
    SOLUTION: A plurality of instances having a labelled order tree structure is inputted, an inner product is calculated on the basis of the structure, and the classification learning of the instances is performed by use of the calculation result of the inner product. A dynamic planning method is applied to a node that is not the terminal end of the labelled order tree on the basis of the corresponding relation in which the order of each node is stored, and the sum of matching related to offspring nodes is calculated.
    COPYRIGHT: (C)2003,JPO

    Abstract translation: 要解决的问题：通过核心方法实现对半结构化数据的分类到标记的顺序树。 输入具有标号的订单树结构的多个实例，根据该结构计算内积，并且通过使用内积的计算结果来执行实例的分类学习。 基于存储每个节点的顺序的对应关系，并且计算与后代节点相关的匹配之和，将动态规划方法应用于不是标记订单树的终端的节点。 版权所有（C）2003，JPO

    DEVICE, SYSTEM AND METHOD FOR RETRIEVING DATABASE, PROGRAM AND STORAGE MEDIUM
    6.
    发明专利
    DEVICE, SYSTEM AND METHOD FOR RETRIEVING DATABASE, PROGRAM AND STORAGE MEDIUM 失效

    公开(公告)号：JP2002351908A

    公开(公告)日：2002-12-06

    申请号：JP2001145684

    申请日：2001-05-15

    Applicant: IBM

    Inventor： SHIBUYA TETSURO ,  KAJIMA HISATSUGU

    IPC: G06F17/30 ,  G06F19/00 ,  G06F19/22 ,  G06F21/00

    Abstract: PROBLEM TO BE SOLVED: To provide a method for retrieving a database in which no request for special processing is needed to a database side and query contents are kept unknown to a database owner and to a person who observes the database retrieval in a network. SOLUTION: The system comprises an array processing part 21 which divides a retrieval array (a retrieval target) and creates a plurality of subarrays in a client 20 which accesses a database server 10 having an accumulation of the array pattern and queries whether the array pattern accumulated in the database server 10 is present in predetermined arrays and a query issuing part 22 which issues a query to the database server 10 with each of the plurality of the subarrays created by the array processing part 21 as a query array.

    Link prediction system, method, and program
    7.
    发明专利
    Link prediction system, method, and program 有权
    链接预测系统，方法和程序

    公开(公告)号：JP2010250377A

    公开(公告)日：2010-11-04

    申请号：JP2009096248

    申请日：2009-04-10

    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： RUDY RAYMOND HARRY PUTRA ,  KAJIMA HISATSUGU

    IPC: G06Q10/04

    Abstract: PROBLEM TO BE SOLVED: To provide a scalable link prediction technology that can cope with the number of dozens to millions nodes.
    SOLUTION: At first, similarity matrices W
    Z , W
    Y , W
    X , are low-rank approximated by a technology such as incomplete Cholesky decomposition. Then, the eigenvalue decomposition of low-rank approximate matrices of the similarity matrices W
    Z , W
    Y , W
    X is performed. Schematically, low-rank approximation is the approximation of one matrix by a product of two rectangular matrices. Here, low-rank approximation facilitates the calculation of eigenvalue decomposition. In the next step, eigenvalues of obtained low-rank approximate matrices of W
    Z , W
    Y , W
    X are used to constitute normalized Laplacian L. Since the normalized Laplacian L is obtained in this manner, V~
    Z , V~
    Y , V~
    X as matrices with respective eigenvectors of low-rank approximate matrices of W
    Z , W
    Y , W
    X arranged therein and L are used to favorably calculate the inverse matrix of the part of (σL+I). When the inverse matrix of (σL+I) is obtained, F can be calculated due to vec(F)=(σL+I)
    -1 vec(F
    * ).
    COPYRIGHT: (C)2011,JPO&INPIT

    Abstract translation: 要解决的问题：提供可以应对数十百万个节点数量的可伸缩链路预测技术。 解决方案：首先，相似度矩阵W Z ，Y ，W X 通过诸如 不完全Cholesky分解。 然后，执行相似矩阵W Z ，W Y ，W X 的低阶近似矩阵的特征值分解。 示意地，低阶近似是由两个矩形矩阵的乘积的一个矩阵的近似。 这里，低阶近似有助于特征值分解的计算。 在下一步骤中，使用获得的W Z ，W Y ，W X 的低等级近似矩阵的特征值来构成归一化拉普拉斯算子L 由于以这种方式获得归一化拉普拉斯算子L，所以V = Z ，V = Y ，V = X 布置在其中的W Z ，W Y ，W X 的低等级近似矩阵用于有利地计算部分的逆矩阵 （σL+ I）。 当获得（σL+ I）的逆矩阵时，可以由于vec（F）=（σL+ I） -1 vec（F * ）计算F 。 版权所有（C）2011，JPO＆INPIT

    METHOD/SYSTEM FOR AUCTION AND RECORDING MEDIUM
    10.
    发明专利
    METHOD/SYSTEM FOR AUCTION AND RECORDING MEDIUM 失效

    公开(公告)号：JP2002163486A

    公开(公告)日：2002-06-07

    申请号：JP2000347349

    申请日：2000-11-14

    Applicant: IBM

    Inventor： KAJIMA HISATSUGU ,  KAJINAGA YASUMASA

    IPC: G06F19/00 ,  G06Q10/04 ,  G06Q30/06 ,  G06Q30/08 ,  G06Q50/00 ,  H04K1/00 ,  H04L9/00 ,  G06F17/60

    Abstract: PROBLEM TO BE SOLVED: To give a solution of the problem of deciding the successful bidder of such a combination auction where bidding of large amount of transaction object (merchandise) is performed when there are many kinds of and a large amount of stocks. SOLUTION: After-mentioned two kinds of restrictions are given to the combination of allowable merchandise. 1. Single bidding can be performed with respect to only one kind of merchandise. 2. The involution relation of the combination of biddable merchandise has hierarchical structure. The optimal partial set of biddings is selected from the set of the biddings restricted like this by applying a dynamic programming.

Patent Agency Ranking