专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明专利

JP2012221431A Related word extraction device, related word extraction method, and related word extraction program 有权
标题翻译：相关词提取装置，相关词提取方法和相关词提取程序
公开(公告)号：JP2012221431A
公开(公告)日：2012-11-12
申请号：JP2011089567
申请日：2011-04-13
申请人： Nippon Telegr & Teleph Corp , 日本電信電話株式会社
发明人： ADACHI TAKAYUKI , UCHIYAMA TOSHIRO , FUJIMURA TAKASHI
IPC分类号： G06F17/30 , G06F17/27
摘要： PROBLEM TO BE SOLVED: To extract a related word of a specified meaning from an object word with an ambiguous meaning.SOLUTION: A related word extraction device 100 prepares co-occurrence word data from a text collection, obtains and groups words that co-occur with the word for respective prescribed words using the co-occurrence word data, extracts group data for an object word 400 from word group data, extracts group data for each support word described in a support word list 500 from the word group data, obtains a co-occurrence frequency of a co-occurrence word (support co-occurrence word) belonging to the group deeply related to the support word and the support word from the support word list 500 and support word group data, sums up the co-occurrence frequency for all the support words to define support co-occurrence word data, obtains a support degree of the support co-occurrence word matched with the co-occurrence word belonging to each related word group of the object word from related word group data and support co-occurrence word data, performs sum-up for each related word group of the object word, and selects the co-occurrence word deeply related to the support co-occurrence word as a related word.
摘要翻译：要解决的问题：从具有含糊含义的对象词中提取具体含义的相关单词。解决方案：相关词提取装置100从文本收集中准备同现字数据，使用同现字数据获得和组合与各个规定词的词共同出现的词，提取用于对象词400从单词组数据中，从单词组数据中提取在支持单词列表500中描述的每个支持词的组数据，获得属于该组词数据的同现词（支持同位词）的同现频率与支持单词列表500的支持单词和支持词深度相关，并且支持单词组数据，总结所有支持词的同现频率以定义支持同时字数据，获得支持词列表的支持度支持与来自相关字组数据和支持同现字数据的对象词的每个相关字组的同现字匹配的同现字，对对象的每个相关单词组进行总结 ct字，并选择与支持共现词深度相关的同现字作为相关词。版权所有（C）2013，JPO＆INPIT

2. 发明专利

JP2012083914A Electronic book analyzer, method, and program 有权
标题翻译：电子书分析仪，方法和程序
公开(公告)号：JP2012083914A
公开(公告)日：2012-04-26
申请号：JP2010228876
申请日：2010-10-08
申请人： Nippon Telegr & Teleph Corp , 日本電信電話株式会社
发明人： MIYATA AKIHIRO , YAMAMOTO CHIHIRO , FUJIMURA TAKASHI
IPC分类号： G06F17/30 , G06Q50/04
CPC分类号： Y02P90/30
摘要： PROBLEM TO BE SOLVED: To efficiently determine a significance of each page of a large selection of electronic books.SOLUTION: Viewing behavior information of an electronic book viewer is collected from a client terminal, and viewing behavior information is totaled. Based on a time length spent on each page for viewing, a significance of that page is determined, and the significance determination result is output. The page significance is determined, not only by the length of viewing time but also by either one or a combined values of the followings: the number of times viewed according to the viewing behavior information (page history), total viewing time, and a value obtained by dividing total viewing time by the number of characters.
摘要翻译：要解决的问题：有效地确定大量电子书每页的意义。
解决方案：从客户终端收集电子书阅览器的行为信息，并查看行为信息。根据花费在每个页面上观看的时间长度，确定该页面的重要性，并输出重要性确定结果。确定页面有效性不仅取决于观看时间的长度，还可以通过以下任意一个或组合的值来确定：根据观看行为信息（页面历史），总观看时间和值的观看次数通过将总观看时间除以字符数获得。版权所有（C）2012，JPO＆INPIT

3. 发明专利

JP2012003729A Work recommendation device and method and program 有权
标题翻译：工作建议设备及方法与程序
公开(公告)号：JP2012003729A
公开(公告)日：2012-01-05
申请号：JP2010141088
申请日：2010-06-21
申请人： Nippon Telegr & Teleph Corp , 日本電信電話株式会社
发明人： KOYA YU , FUJIMURA TAKASHI
IPC分类号： G06F17/30 , H04N7/173
摘要： PROBLEM TO BE SOLVED: To appropriately recommend a work such as a program in a video distribution service using the Internet line.SOLUTION: A work recommendation device acquires work information and encyclopedia on a WWW, acquires pertinent items in the encyclopedia on the WWW based on the title of a work in the work information, calculates similarity between works from the similarity of the pertinent items in the encyclopedia on the WWW about the combination of all the two works in the work information, acquires the viewing history of each user with respect to the work, calculates the recommendation scores of the work with respect to the user by using the viewing history of the user and the work similarity about the user from the acquired viewing history and all the works whose information is stored in work information storage means, acquires recommendation scores stored in the recommendation score storage means with respect to each work, and selects the identifiers of the high rank N items of works of the recommendation scores as recommendation works.
摘要翻译：要解决的问题：适当地推荐使用互联网线路的视频分发服务中的程序等工作。解决方案：工作推荐设备在万维网上获取工作信息和百科全书，根据工作信息中的作品标题，获取万维网百科全书中的有关项目，根据相关项目的相似性计算作品之间的相似性在WWW的百科全书中关于工作信息中的所有两个作品的组合，获取每个用户对作品的观看历史，通过使用观看历史来计算相对于用户的作品的推荐分数从所获取的观看历史中的用户和用户的工作相似性以及信息存储在工作信息存储装置中的所有作品获取关于每个作品存储在推荐分数存储装置中的推荐分数，并且选择高等级N项作品推荐评分作为推荐作品。版权所有（C）2012，JPO＆INPIT

4. 发明专利

JP2011165092A Providing device and acquisition system of document image relevant information 有权
标题翻译：提供文件图像相关信息的设备和获取系统
公开(公告)号：JP2011165092A
公开(公告)日：2011-08-25
申请号：JP2010029521
申请日：2010-02-12
申请人： Nippon Telegr & Teleph Corp , 日本電信電話株式会社
发明人： SHIOBARA TOSHIKO , FUJIMURA TAKASHI , MIYATA AKIHIRO
IPC分类号： G06F17/30 , H04N1/00
摘要： PROBLEM TO BE SOLVED: To allow a user to easily obtain information relevant to a word described in a printed book or the like through a terminal or the like carried thereby.
SOLUTION: The document image relevant information providing device includes: a receiving means which receives image data including a character string image from a user terminal connected to a communication network; a character recognition means which performs character recognition processing to the received image data to extract a character string from the image data; a word information extraction means which analyzes the character string extracted by the character recognition means to extract a word group from the character string; an importance determination means which selects an important keyword with high level of importance from the word group extracted by the word information extraction means based on a predetermined condition; a relevant information generation means which generates relevant information including link information to the information related to the important keyword selected by the importance determination means; and a transmitting means which transmits the relevant information generated by the relevant information generation means to the user terminal.
COPYRIGHT: (C)2011,JPO&INPIT
摘要翻译：要解决的问题：为了允许用户通过由其承载的终端等容易地获得与打印的书籍等中描述的词相关的信息。文件图像相关信息提供装置包括：接收装置，从连接到通信网络的用户终端接收包括字符串图像的图像数据; 字符识别装置，对所接收的图像数据执行字符识别处理，以从图像数据提取字符串; 字信息提取装置，分析由字符识别装置提取的字符串，以从字符串中提取字组; 基于预定条件从字信息提取装置提取的字组中选择具有高重要度的重要关键词的重要性确定装置; 相关信息生成装置，生成包括与重要性确定装置选择的重要关键字有关的信息的链接信息的相关信息; 以及发送装置，其将由相关信息生成装置生成的相关信息发送给用户终端。版权所有（C）2011，JPO＆INPIT

5. 发明专利

JP2011129070A Retrieval device, method, and program 有权
标题翻译：检索设备，方法和程序
公开(公告)号：JP2011129070A
公开(公告)日：2011-06-30
申请号：JP2009289787
申请日：2009-12-21
申请人： Nippon Telegr & Teleph Corp , 日本電信電話株式会社
发明人： MIYATA AKIHIRO , FUJIMURA TAKASHI
IPC分类号： G06F17/30 , G06T1/00
摘要： PROBLEM TO BE SOLVED: To reduce a calculation cost required for matching by improving a probability of obtaining a correct retrieval result when giving an index to a photographic object.
SOLUTION: A retrieval device specifies a character area in an input document, specifies a peripheral area of each character area, determines an area different from a background area in the peripheral area as a partial image area, and outputs the character area, the partial image area around the character area, and an appearance position in the document in which a character string appears, in relation to one another to an index DB. The retrieval device specifies a character area in an input image for retrieval, specifies a peripheral area of each character area, compares each peripheral area with the background area of the image for retrieval, determines an area different from the background area in the peripheral area as a partial image area, inquires of the index DB about a document or a position in the document in which the character area and the partial image area around the character area appear, and outputs the result of the inquiry.
COPYRIGHT: (C)2011,JPO&INPIT
摘要翻译：要解决的问题：为了在为摄影对象给出索引时通过提高获得正确的检索结果的概率来减少匹配所需的计算成本。解决方案：检索装置指定输入文档中的字符区域，指定每个字符区域的外围区域，确定与外围区域中的背景区域不同的区域作为部分图像区域，并输出字符区域，字符区域周围的部分图像区域和文档中出现字符串的外观位置相对于索引DB。检索装置指定用于检索的输入图像中的字符区域，指定每个字符区域的外围区域，将每个外围区域与用于检索的图像的背景区域进行比较，确定与周边区域中的背景区域不同的区域作为部分图像区域，询问索引DB关于文档或文档中位置，其中出现字符区域周围的字符区域和部分图像区域的文档或位置，并输出查询结果。版权所有（C）2011，JPO＆INPIT

6. 发明专利

JP2010250496A Space-time retrieval device, method, and program 有权
标题翻译：空间检索设备，方法和程序
公开(公告)号：JP2010250496A
公开(公告)日：2010-11-04
申请号：JP2009098279
申请日：2009-04-14
申请人： Nippon Telegr & Teleph Corp , 日本電信電話株式会社
发明人： KURASHIMA TAKESHI , FUJIMURA TAKASHI
IPC分类号： G06F19/00 , G06F17/30 , G06Q50/00 , G06Q50/10
摘要： PROBLEM TO BE SOLVED: To search an optimal space-time region for most satisfactorily characterizing persons who have a certain feeling by using personal delivery information such as a blog as a target. SOLUTION: A space-time retrieval device is configured to input analysis object data, constraints to a searched space-time region and conditions to characterize persons, and to determine whether or not the content of the text of the analysis object data is written by the persons pertinent to the conditions to characterize the persons by using the conditions as a target value for searching for the optimal space-time region, and to generate transaction having the determination result and location information and time information given to the analysis object data as elements, and to derive space-time region conditions to maximize the appearance probability of the transaction pertinent to the target value from the group by extracting a numerical attribute correlation rule. COPYRIGHT: (C)2011,JPO&INPIT
摘要翻译：要解决的问题：通过使用诸如博客的个人交付信息作为目标，来搜索具有某种感觉的人的最令人满意的人物的最佳时空区域。解决方案：空时检索装置被配置为输入分析对象数据，对所搜索的时空区域的约束和用于表征人的条件，并且确定分析对象数据的文本的内容是否为通过使用条件作为搜索最佳时空区域的目标值的与条件相关的人员写入的事物，并且生成具有给分析对象数据的确定结果和位置信息和时间信息的交易作为元素，并且通过提取数字属性相关规则来导出空间 - 时间区域条件以最大化与组中的目标值相关的事务的出现概率。版权所有（C）2011，JPO＆INPIT

7. 发明专利

JP2010204864A Information processing apparatus and method and program for detecting commodity solving complaint 有权
标题翻译：信息处理设备和检测商品解决方案的方法和程序
公开(公告)号：JP2010204864A
公开(公告)日：2010-09-16
申请号：JP2009048548
申请日：2009-03-02
申请人： Nippon Telegr & Teleph Corp , 日本電信電話株式会社
发明人： SAKAI TOSHIYUKI , FUJIMURA TAKASHI
IPC分类号： G06F19/00 , G06F17/30 , G06Q10/00 , G06Q30/02 , G06Q50/00
摘要： PROBLEM TO BE SOLVED: To discern and extract complaints about demerits or unprofitable events from various complaints as complaints usable in the recommendation of a commodity, etc., and to extract a semantic relationship between a complaint and a commodity solving the complaint, from relationships between complaints and commodities. SOLUTION: A document is obtained from a document set database and improving actions are extracted and stored in an improving action storage means. The improving actions are obtained from the improving action storage means. Of the subjects of the improving actions, negative ones are regarded as complaints. A group of documents is obtained from the document set database and the name of a commodity is obtained from a commodity database. From the group of documents, the co-occurrence frequency of a combination of three, that is, the name of the commodity, a complaint, and an improving action, is determined; it is determined that the higher the co-occurrence frequency, the higher the relationship between the commodity and the complaint will be. The co-occurrence frequency of the combination of three, the name of the commodity, the complaint, and the improving action, is stored in a relationship storage means so as to determine the relationship between the complaint and the commodity. COPYRIGHT: (C)2010,JPO&INPIT
摘要翻译：要解决的问题：从各种投诉中辨认和提取关于违规或无利可图事件的投诉，作为商品建议中可用的投诉等，并提取投诉和解决投诉的商品之间的语义关系，从投诉和商品之间的关系。解决方案：从文件集数据库获得文档，提取并提取改进动作并将其存储在改进动作存储装置中。从改进动作存储装置获得改善动作。在改善行动的主题中，消极的被认为是投诉。从文件集数据库中获取一组文档，并从商品数据库获取商品的名称。从一组文件中，确定了三种组合的共同发生频率，即商品名称，投诉和改进措施; 确定共同发生频率越高，商品与投诉之间的关系越高。三种组合的共现频率，商品名称，投诉和改进措施，存储在关系存储装置中，以确定投诉与商品之间的关系。版权所有（C）2010，JPO＆INPIT

8. 发明专利

JP2009277002A Experience information analysis device, experience information analysis method and program 有权
标题翻译：体验信息分析设备，体验信息分析方法与程序
公开(公告)号：JP2009277002A
公开(公告)日：2009-11-26
申请号：JP2008127406
申请日：2008-05-14
申请人： Nippon Telegr & Teleph Corp , 日本電信電話株式会社
发明人： KURASHIMA TAKESHI , FUJIMURA TAKASHI
IPC分类号： G06F19/00
摘要： PROBLEM TO BE SOLVED: To analyze a tendency of a time series of people related to experience with a set of data such as a blog wherein a history of experience of an individual is described in a natural language and wherein information allowing unique identification of a date or a person having described it or having actually performed the experience described therein is imparted thereto as a target. SOLUTION: This experience information analysis device analyzes the analysis target data, extracts notation of elements constituting experience information of the person, generates transaction comprising the information allowing the unique identification of the person, the information allowing the unique identification of the date, and the notation of the extracted experience elements, generates a data sequence arranged with the notation of the experience elements in each person from the generated transaction, extracts a sequential pattern representing arrangement of the time series of the notation of the experience elements from the data sequence, and generates a sequential rule for predicting the next appearing notation of the experience element from some arrangement of the notation of the experience elements based on a set of the sequential pattern. COPYRIGHT: (C)2010,JPO&INPIT
摘要翻译：要解决的问题：通过一组数据来分析与经验有关的时间序列的趋势，例如博客，其中以自然语言描述个人的经验历史，并且其中允许唯一标识的信息将其描述或实际执行其中所描述的经历的日期或人员作为目标。解决方案：该体验信息分析装置分析分析目标数据，提取构成该人的体验信息的元素的符号，生成包含允许人的唯一识别的信息的交易，允许唯一识别日期的信息，以及提取的经验元素的符号，从生成的事务生成以每个人的体验元素的符号排列的数据序列，从数据序列中提取表示体验元素的符号的时间序列的排列的顺序模式，并且基于一组序列模式来生成用于从经验元素的符号的一些排列中预测经验元素的下一个出现符号的顺序规则。版权所有（C）2010，JPO＆INPIT

9. 发明专利

JP2008310626A Automatic tag impartment device, automatic tag impartment method, automatic tag impartment program and recording medium recording the program 有权
标题翻译：自动标签封装设备，自动标签封装方法，自动标签封装程序和记录介质记录程序
公开(公告)号：JP2008310626A
公开(公告)日：2008-12-25
申请号：JP2007158422
申请日：2007-06-15
申请人： Nippon Telegr & Teleph Corp , 日本電信電話株式会社
发明人： FUJIMURA SHIGERU , FUJIMURA TAKASHI
IPC分类号： G06F17/30
摘要： PROBLEM TO BE SOLVED: To provide an automatic tag impartment device capable of supporting a browser by imparting a proper tag to a document and arranging the document, and capable of imparting the tag with high accuracy.
SOLUTION: This automatic tag impartment device has: a document collection means 1 collecting the tag-equipped document from the Internet 2, and storing it into a tag-equipped document DB 3; a tag integration means 4 integrating the tags showing the same meaning based on a characteristic term inside a document set set with the tag to each the tag inside the DB 3, and storing it into a tag information DB 5; a tag importance decision means 6 deciding whether the tag shows a specific subject or not based on the characteristic term to each the tag inside the DB 3, and storing it into the tag information DB 5; a tag impartment means 7 imparting the tag to the input document based on a decision result of the tag importance decision means 6 and an integration result of the tag integration means 4; and a data careful selection means 8 setting each the document inside the DB 3 as input, and removing the document imparted with the tag unrelated to the document from the DB 3.
COPYRIGHT: (C)2009,JPO&INPIT
摘要翻译：要解决的问题：提供一种能够通过向文件赋予适当的标签并布置文档并且能够高精度地赋予标签来支持浏览器的自动标签隔离装置。解决方案：该自动标签分配装置具有：文件收集装置1，从互联网2收集装有标签的文档，并将其存储到装有标签的文档DB 3中; 标签集成装置4，将根据具有标签的文档集合中的特征项显示相同含义的标签集成到DB 3内的每个标签，并将其存储到标签信息DB 5中; 标签重要性判定单元6根据特征项判定标签是否显示特定对象，并将其存储到标签信息DB5中; 标签附加装置7根据标签重要性决定装置6的判定结果和标签整合装置4的积分结果，将标签赋予输入文档; 并且数据仔细选择装置8将DB 3内的每个文档设置为输入，并且从DB 3中删除被赋予与该文档无关的标签的文档。（C）2009，JPO＆INPIT

10. 发明专利

JP2008134941A Inter-metadata similarity measuring apparatus and metadata hierarchizing apparatus, inter-metadata similarity measuring method and metadata hierarchizing method, inter-metadata similarity measuring program and metadata hierarchizing program, and recording medium recording these programs 有权
标题翻译：跨元数据相似度测量装置和元数据分层设备，元数据类似度测量方法和元数据分层方法，元数据类似度量测程序和元数据分层程序以及记录介质记录这些程序
公开(公告)号：JP2008134941A
公开(公告)日：2008-06-12
申请号：JP2006322050
申请日：2006-11-29
申请人： Nippon Telegr & Teleph Corp , 日本電信電話株式会社
发明人： FUJIMURA SHIGERU , FUJIMURA TAKASHI
IPC分类号： G06F17/30
摘要： PROBLEM TO BE SOLVED: To easily set a threshold in an apparatus for measuring similarity between metadata attached to a document.
SOLUTION: The inter-metadata similarity measuring apparatus is provided with a document retrieval means 11 for retrieving a document to which specific metadata are attached from a metadata-attached document DB (database) 14, a feature value extraction means 12 for extracting a vector in which a pair of scores expressing relationship between words/phrases in the document and the contents of the document of the words/phrases is used as an element from a document group retrieved by the document retrieval means 11 as the feature value of the metadata, and a similarity calculation means 13 for measuring similarity between target metadata by using a metadata feature value DB (database) 15 storing obtained feature values.
COPYRIGHT: (C)2008,JPO&INPIT
摘要翻译：要解决的问题：在用于测量附加到文档的元数据之间的相似性的装置中容易地设置阈值。解决方案：元数据之间相似度测量装置具有文件检索装置11，用于从附于元数据的文档DB（数据库）14检索附加有特定元数据的文档，特征值提取装置12，用于提取将文档中的单词/短语之间的关系的一对分数和单词/短语的文档的内容用作从文档检索装置11检索到的文档组中的元素作为元素的特征值的向量元数据和相似度计算装置13，用于通过使用存储获得的特征值的元数据特征值DB（数据库）15来测量目标元数据之间的相似度。版权所有（C）2008，JPO＆INPIT

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式