公开(公告)号：JP2012190080A

公开(公告)日：2012-10-04

申请号：JP2011050830

申请日：2011-03-08

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International Business Maschines Corporation

Inventor： RUDY RAYMOND HARRY PUTRA ,  TSUBOI YUTA ,  NASUKAWA TETSUYA ,  CHANG YAO-CHUNG

IPC: G06F17/27 ,  G06F17/28

CPC classification number: G06F17/2809

Abstract: PROBLEM TO BE SOLVED: To provide a technique that makes it possible to list a set of corresponding terms in a different language using a random walk technique with a reasonable calculation amount even when a large number of terms are targeted.SOLUTION: The method comprises the steps of: creating a technical term set of a first language (e.g., Japanese), a general term set of the first language, a technical term set of a second language (e.g., English), and a general term set of the second language; creating a bipartite graph for each of the first and second languages by connecting the technical term set and the general term set of the language with links based on corpus information; creating links between the general terms using a translation dictionary associating general terms between the two languages (e.g., an English-Japanese dictionary) to create a bipartite graph connecting the general term sets of the two languages; creating an association matrix M based on weighted information of links of the bipartite graphs within each language and between the two languages; and calculating a similarity matrix Q defined by Q=(1-c)(I-cM).

Abstract translation: 要解决的问题：提供一种使得可以使用具有合理计算量的随机游走技术以不同语言列出一组相应术语的技术，即使当针对大量术语时也是如此。 解决方案：该方法包括以下步骤：创建第一语言（例如，日语）的技术术语集，第一语言的通用术语集，第二语言的技术术语集（例如英语）， 和第二语言的一般术语集; 通过将技术术语集和语言的一般术语集合与基于语料库信息的链接相连接来为每个第一和第二语言创建二分图; 使用与两种语言之间的一般术语（例如，英语 - 日语词典）相关联的翻译字典在一般术语之间建立链接，以创建连接两种语言的通用术语集合的二分图; 基于每种语言和两种语言之间的二分图的链接的加权信息来创建关联矩阵M; 并且计算由Q =（1-c）（I-cM） 1定义的相似度矩阵Q。 版权所有（C）2013，JPO＆INPIT

公开(公告)号：JP2009223745A

公开(公告)日：2009-10-01

申请号：JP2008069112

申请日：2008-03-18

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： NASUKAWA TETSUYA ,  NAKAMURA HIROYOSHI ,  TAKUMA DAISUKE ,  SUZUKI AKIKO

IPC: G06F13/00

Abstract: PROBLEM TO BE SOLVED: To provide a technology for acquiring evaluation information of an object in a virtual-reality space. SOLUTION: An evaluation acquisition apparatus includes: a first conversation data acquiring section for acquiring neighboring conversation data composed of conversations which are made around an object of interest to be evaluated; a second conversation data acquiring section for acquiring wide-range conversation data composed of conversations made in an area in a virtual-reality space wider than a neighboring area of the object of interest; and an acquiring section for specifying expressions frequently appearing in the neighboring conversation data by use of the neighboring conversation data and the wide-range conversation data to acquire the expressions as the evaluation information of the object of interest. COPYRIGHT: (C)2010,JPO&INPIT

Abstract translation: 要解决的问题：提供一种用于获取虚拟现实空间中的对象的评估信息的技术。 评估获取装置包括：第一对话数据获取部分，用于获取由待评估的关注对象周围的对话组成的邻近对话数据; 第二会话数据获取部分，用于获取由比感兴趣对象的相邻区域更宽的虚拟现实空间中的区域中进行的对话组成的宽范围对话数据; 以及获取部分，用于通过使用相邻会话数据和广范围对话数据来指定频繁出现在相邻会话数据中的表达，以获取表达式作为感兴趣对象的评估信息。 版权所有（C）2010，JPO＆INPIT

公开(公告)号：JP2008176721A

公开(公告)日：2008-07-31

申请号：JP2007011633

申请日：2007-01-22

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： NAKAMURA HIDEFUMI ,  MURAKAMI AKIKO ,  NASUKAWA TETSUYA

IPC: G06F17/30 ,  G06F19/00

CPC classification number: G06F17/30864

Abstract: PROBLEM TO BE SOLVED: To enable a user to automatically search a message which matches his or her interest and includes a speech of high importance, within a discussion thread. SOLUTION: Degrees in which preliminarily registered words expressing approval, opposition, criticism, consent, etc. are included in contents of each message are expressed numerically. A positive point is given to a message to which a reply of a message including agreeable and positive words is received, and a negative point is given to a message to which a reply of a message including critical and negative words is received. A positive point is given to a message including words matching user's interest. These points are totalized to calculate the total degree of importance of each message, and messages having high total degrees of importance are determined to match user's interest. COPYRIGHT: (C)2008,JPO&INPIT

Abstract translation: 要解决的问题：为了使用户能够在讨论线程内自动搜索与他或她的兴趣相匹配的消息并且包括高度重要的语音。

解决方案：表达批准，反对，批评，同意等的初步注册单词包含在每条消息的内容中的程度以数字表示。 给出一个消息，其中接收到包括合意和正确的单词的消息的回复的消息，并且给出了消息的负点，消息包括包括关键和否定单词的消息的回复。 给出包含与用户兴趣相匹配的单词的消息的积极点。 总计这些点以计算每个消息的总重要程度，确定具有高总重要度的消息以匹配用户的兴趣。 版权所有（C）2008，JPO＆INPIT

公开(公告)号：CA2451083A1

公开(公告)日：2003-02-13

申请号：CA2451083

申请日：2002-07-19

Applicant: IBM

Inventor： NASUKAWA TETSUYA ,  MATSUZAWA HIROFUMI ,  MURAKAMI AKIKO

IPC: G06F17/27 ,  G06F17/30 ,  G06F40/00 ,  G06F17/28

Abstract: A supporting system for efficiently creating a synonym candidate when a thesaurus used in text mining is compiled and a method for creating a synony m candidate are disclosed. A synonym candidate acquiring device (130) creates, for each author, an author synonym candidate set containing synonym candidat es similar to an inputted word from data (110) on the author and creates a whol e synonym candidate set containing synonym candidates similar to the inputted word from the whole data (120). A synonym candidate judging device (150) evaluates the synonym candidates of the whole data (120) on receiving the created synonym candidate set (140). During the evaluation, a status, ~absolute~, is added to a word agreeing with the word rating as the first place in the synonym candidates for each author; and a status, ~negative~, i s added to a word agreeing with the word rating as the second or later place.

公开(公告)号：DE112012005344T5

公开(公告)日：2014-08-28

申请号：DE112012005344

申请日：2012-11-16

Applicant: IBM

Inventor： IKAWA YOHEI ,  ENOKI MIKI ,  RUDY RAYMOND HARRY PUTRA ,  NASUKAWA TETSUYA

IPC: G06F17/30

Abstract: Problem: Stoßweise übertragene Daten (Daten, die in einem vom Üblichen abweichenden Umfang verbreitet werden) in einem sozialen Medium sollen schneller und genauer erkannt werden. Mittel zur Lösung: Es wird ein Verfahren zum Auswählen einer erkannten Gruppe von Nachrichten aus der Vielzahl von Nachrichten vorgeschlagen, die durch eine Vielzahl von Benutzern in einem sozialen Medium gesendet wurden. Innerhalb der Nachrichten sind Nachrichten enthalten, die durch Benutzer gesendet wurden, die von dem Benutzer der ursprünglichen Nachricht verschieden sind, und diese Nachrichten zitieren die durch den Benutzer der ursprünglichen Nachricht gesendete Nachricht. Für die Nachrichten, die durch Benutzer gesendet wurden, die von dem Benutzer der ursprünglichen Nachricht verschieden sind, die die durch den Benutzer der ursprünglichen Nachricht gesendete Nachricht zitieren, werden bereitgestellt: ein Schritt zum Durchführen einer Erstbewertung des Potenzials für das erneute Weitersenden; ein Schritt zum Ermitteln eines einzelnen Benutzers, der eine Nachricht gesendet hat, für die das Ergebnis der Erstbewertung einen vorgegebenen ersten Schwellenwert überschritten hat; einen Schritt zum Ermitteln einer Gruppe von Nachrichten, die einer Gruppe weitergesendeter Nachrichten ähnlich ist, auf der Grundlage einer Gruppe weitergesendeter Nachrichten, die die durch den einzelnen Benutzer gesendete Nachricht zitieren; einen Schritt zum Durchführen einer Zweitbewertung des Potenzials, dass die ähnliche Gruppe von Nachrichten erneut weitergesendet wird; und einen Schritt zum Auswählen einer Gruppe von Nachrichten, für die das Ergebnis der Zweitbewertung einen vorgegebenen zweiten Schwellenwert überschritten hat.

公开(公告)号：JP2008269231A

公开(公告)日：2008-11-06

申请号：JP2007110559

申请日：2007-04-19

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： KANEYAMA HIROSHI ,  NISHIYAMA RISA ,  WATANABE HIDEO ,  NASUKAWA TETSUYA

IPC: G06Q30/02 ,  G06Q30/06 ,  G06Q50/00

CPC classification number: G06Q30/02 ,  G06Q30/0256 ,  G06Q30/0277

Abstract: PROBLEM TO BE SOLVED: To substantially prevent a negative image from being caused in one or both of the transmission origin of an article and an adviser, or to prevent feeling of incompatibility from being given to a reader in a system for selecting advertisements with respect to the content of an article on a Web page. SOLUTION: The appearance of advertisements giving any feeling of incompatibility to the reader of a Web page or a blog, and the interest label of any of "assailant", "victim" and "assessment" is systematically applied to any pertinent one of such things as an enterprise, person and product mentioned in the article. Then, an advertisement selecting mechanism is controlled so that the appearance of advertisements associated with the interest label can be suppressed. The advertisement selection mechanism to be used for this purpose is able to calculate the affinity or relevancy of the article and the advertisement as numerics. Therefore, the suppression of advertisement is achieved by subtracting the numerics of the affinity. Also, it is desired that the degree of suppression by the interest label is gradually attenuated according to the time. COPYRIGHT: (C)2009,JPO&INPIT

Abstract translation: 要解决的问题：为了基本上防止在物品和顾问的传送来源中的一个或两个中引起负像，或者防止在用于选择广告的系统中向读者提供不兼容的感觉 关于网页上的文章的内容。 解决方案：对网页或博客的读者感到不兼容的广告的外观，以及“凶手”，“受害者”和“评估”中的任何一个的兴趣标签的系统应用于任何相关的 如文章中提到的企业，人员和产品等。 然后，控制广告选择机制，使得可以抑制与兴趣标签相关联的广告的外观。 用于此目的的广告选择机制能够计算文章和广告作为数字的亲和性或相关性。 因此，通过减去亲和度的数字来实现广告的抑制。 此外，期望利用标签的抑制程度根据时间逐渐衰减。 版权所有（C）2009，JPO＆INPIT

公开(公告)号：JP2008225579A

公开(公告)日：2008-09-25

申请号：JP2007059088

申请日：2007-03-08

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International Business Maschines Corporation

Inventor： TAKEUCHI HIROYOSHI ,  L V SUBRAMANIAM ,  NASUKAWA TETSUYA ,  ROY SHOURYA

IPC: G06F17/30 ,  G06F17/28 ,  G06F19/00

CPC classification number: G06F17/30985 ,  G06Q30/0256

Abstract: PROBLEM TO BE SOLVED: To retrieve a keyword, that decides the occurrence of an event, from a text, in which a process to the occurrence of the event is recorded. SOLUTION: The system includes: a text input part for inputting a plurality of partial texts acquired by dividing each text into a plurality parts by associating them with events occurring after a process recorded in the text; a prediction unit adjusting part for adjusting an event prediction unit to maximize the rate of the texts with which the input event and prediction result are matched among a first text group selected from the plurality of corresponding partial texts with respect to each block; a prediction processing part for generating a prediction result by inputting each of the second text group selected from the plurality of corresponding partial texts to the adjusted event prediction unit with respect to each block; and a retrieval part for comparing the input events and prediction results for every partial text and calculating predicting precision for the second text group of the event prediction unit for every block, and for retrieving a keyword from a block whose predicting precision is higher than the reference value. COPYRIGHT: (C)2008,JPO&INPIT

Abstract translation: 要解决的问题：从记录事件发生的处理的文本中检索决定事件的发生的关键字。 解决方案：该系统包括：文本输入部分，用于输入通过将每个文本分割成多个部分而获取的多个部分文本，将其与在文本中记录的处理之后发生的事件相关联; 预测单元调整部，用于调整事件预测单元，以使从所述多个对应的部分文本中选择的第一文本群相对于每个块最大化输入事件和预测结果匹配的文本的比率; 预测处理部分，用于通过将从多个对应的部分文本中选择的每个第二文本组相对于每个块输入到经调整的事件预测单元来产生预测结果; 以及检索部分，用于比较每个部分文本的输入事件和预测结果，并为每个块计算事件预测单元的第二文本组的预测精度，并且从预测精度高于参考的块中检索关键字 值。 版权所有（C）2008，JPO＆INPIT

公开(公告)号：JP2007179276A

公开(公告)日：2007-07-12

申请号：JP2005376579

申请日：2005-12-27

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： WATANABE HIDEO ,  NASUKAWA TETSUYA

IPC: G06F13/00 ,  G06Q30/02 ,  G06Q30/06 ,  G06Q50/00

CPC classification number: G06Q30/02 ,  G06Q30/0251 ,  G06Q30/0256

Abstract: PROBLEM TO BE SOLVED: To provide a method for judging conformity between a character string and merchandise. SOLUTION: The method for judging conformity between a character string including evaluation and merchandise information including the characteristic description of merchandise stored in a storage device is provided, wherein the character string includes an evaluation object as the object of evaluation and evaluation expressions as the evaluation of the evaluation object. The method includes the steps of: judging, on the basis of the evaluation object and the evaluation expressions, whether the evaluation expressions are positive evaluation for the evaluation object or the evaluation expressions are negative evaluation for the evaluation object; and judging, when the evaluation expressions are positive evaluation for the evaluation object, that merchandise whose evaluation expressions are included in the characteristic description of the merchandise is conforming merchandise and judging that the merchandise which is not included in the characteristic description is non-conforming merchandise. COPYRIGHT: (C)2007,JPO&INPIT

Abstract translation: 要解决的问题：提供一种用于判断字符串与商品之间的一致性的方法。 提供了一种用于判断包括评价的字符串与包含存储在存储装置中的商品的特征描述的商品信息之间的一致性的方法，其中，所述字符串包括作为评估对象的评估对象和评价表达式 评价对象的评价。 该方法包括以下步骤：根据评价对象和评价表达来判断评价对象是评价对象的正评价，评价对象为评价对象的否定评价; 并且当评估表达式对于评估对象的肯定评估时，评估表达式被包括在商品的特征描述中的商品是符合商品并且判断未包括在特征描述中的商品是不合格商品 。 版权所有（C）2007，JPO＆INPIT

公开(公告)号：JP2001075966A

公开(公告)日：2001-03-23

申请号：JP24105799

申请日：1999-08-27

Applicant: IBM

Inventor： NASUKAWA TETSUYA ,  NAGANO TORU

IPC: G06F17/30

Abstract: PROBLEM TO BE SOLVED: To provide a method and a system for acquiring effective knowledge from a lot of data. SOLUTION: Concerning these method and system for acquiring the effective knowledge by extracting a concept having a special feature from a lot of data including documents, this system has a concept extracting device and a characteristics concept extracting device. In the said concept extraction 120, concepts by categories are extracted from data including document data. The said characteristics concept extracting device is a device for extracting a characteristics concept out of the said extracted concepts and concerning the said concepts by categories, a concept wherein the ratio of occupancy in a concept belonging to the other corresponding category exceeds a specified value is extracted out of concepts belonging to the same category.

公开(公告)号：GB2511235A

公开(公告)日：2014-08-27

申请号：GB201409114

申请日：2012-11-16

Applicant: IBM

Inventor： IKAWA YOHEI ,  ENOKI MIKI ,  RUDY RAYMOND HARRY PUTRA ,  NASUKAWA TETSUYA

IPC: G06F17/30 ,  G06Q50/00

Abstract: [Problem] To more accurately, and at an earlier stage, detect burst information (information that is diffused on a different scale than normally) in a social medium. [Solution] Proposed is a method which, in a social medium, selects a group of specific messages from among a plurality of messages transmitted by a plurality of users. Here, in the messages are included messages for which a message that has been transmitted by one user is quoted and retransmitted by another user. The present method is provided with steps which: for messages in which the message that has been transmitted by the one user has been quoted and retransmitted by another user, perform primary evaluation of possibility of further re-transmission; identify the one user who has transmitted a message for which the result of the primary evaluation is higher than a pre-defined first threshold; on the basis of the group of messages for which messages transmitted by the one user have been quoted and retransmitted, identify a group of messages similar to the group of messages that have been retransmitted; perform secondary evaluation of the possibility that the similar group of messages will be further retransmitted; and select a group of messages for which the result of the secondary evaluation is higher than a pre-defined second threshold.