분배규칙 생성을 위한 방법 및 그를 위한 질의 분석 장치
    11.
    发明授权
    분배규칙 생성을 위한 방법 및 그를 위한 질의 분석 장치 有权
    用于生成分布规则的方法和用于查询分析的装置

    公开(公告)号:KR101226161B1

    公开(公告)日:2013-01-24

    申请号:KR1020120083208

    申请日:2012-07-30

    CPC classification number: G06F17/3071

    Abstract: PURPOSE: A method for generating a distribution rule and a query analyzing device therefor are provided to generate the distribution rule for reducing a response speed of the whole system by minimizing the join calculation and a network input/output time between nodes of a distributed parallel system. CONSTITUTION: A pattern extracting/grouping unit(140) extracts a pattern from query sets included in a query list and groups the pattern based on type information in order to generate a type pattern group of a query unit. A grouping unit(150) generates a type pattern group of a service unit based on the type information. A distribution rule generating unit(160) generates a distribution rule by generating a pattern set composition file of the query unit with the type pattern group of the service unit. [Reference numerals] (100) Query analyzing device; (110) Query list DB; (120) Distribution rule DB; (130) Query list extracting unit; (140) Pattern extracting/grouping unit; (150) Grouping unit; (160) Distribution rule generating unit

    Abstract translation: 目的:提供一种生成分发规则的方法及其查询分析设备,以生成分布规则,用于通过最小化连接计算和分布式并行系统的节点之间的网络输入/输出时间来降低整个系统的响应速度 。 构成:模式提取/分组单元(140)从包括在查询列表中的查询集中提取模式,并且基于类型信息对模式进行分组,以便生成查询单元的类型模式组。 分组单元(150)基于类型信息生成服务单元的类型模式组。 分发规则生成单元(160)通过利用服务单元的类型模式组生成查询单元的模式组合文件来生成分配规则。 (附图标记)(100)查询分析装置; (110)查询列表DB; (120)分配规则DB; (130)查询列表提取单元; (140)模式提取/分组单元; (150)分组单位; (160)分布规则生成单元

    하이브리드 규칙 추론 장치 및 그 방법
    15.
    发明授权
    하이브리드 규칙 추론 장치 및 그 방법 有权
    混合规则理由的装置和方法

    公开(公告)号:KR101535807B1

    公开(公告)日:2015-07-10

    申请号:KR1020140153763

    申请日:2014-11-06

    CPC classification number: G06N5/046

    Abstract: 본발명은하이브리드규칙추론장치및 방법에관한것으로, 입력된추론규칙을스키마규칙, 인스턴스규칙, 혼합규칙중 적어도하나의규칙으로분류하는추론규칙분류부, 스키마규칙을기 저장된스키마트리플에적용하여추론하고, 혼합규칙을스키마트리플에적용하여인스턴스규칙으로구체화하는메모리기반추론부, 상기메모리기반추론부에서구체화된인스턴스규칙또는상기추론규칙분류부에서분류된인스턴스규칙을기 저장된인스턴스트리플에적용하여추론하는디스크기반분산/병렬추론부를포함한다.

    Abstract translation: 本发明涉及一种用于推导混合规则的装置和方法。 该设备包括:推理规则分类单元,用于将输入的推理规则分类为模式规则,实例规则和混合规则中的至少一个; 基于存储器的推理单元,用于通过将模式规则应用于预存储的模式三进制来执行推理,并将混合规则应用于将被规定为实例规则的模式三元组; 以及基于盘的分配/并行推理单元,用于通过将基于存储器的推理单元中指定的实例规则或从推理规则分类单元分类的实例规则应用于预存储的实例三进制来执行推理。

    맵리듀스 기반 분산 병렬 개체 추출 시스템 및 방법
    16.
    发明授权
    맵리듀스 기반 분산 병렬 개체 추출 시스템 및 방법 有权
    基于MAPREDUCE提取分布式并发实体的系统和方法

    公开(公告)号:KR101255060B1

    公开(公告)日:2013-04-16

    申请号:KR1020120077379

    申请日:2012-07-16

    Abstract: PURPOSE: A MapReduce based dispersion parallel entity extracting system and a method thereof are provided to guarantee shortened entity extracting response time by extracting entity based on a MapReduce framework. CONSTITUTION: A master server device(100) distributes target document data to slave server devices(200a-200N) by dividing an input document into the target document data. The slave server device converts the target document data into a data format which is able to be processed in a MapReduce framework, divides the content of the converted document into sentences, and divides the divided sentences into construction units. The slave server device extracts the combination of the construction units as entity candidates and defines a relationship between the extracted entities. [Reference numerals] (100) Master server; (200a) Slave server 1; (200b) Slave server 2; (200N) Slave server N;

    Abstract translation: 目的:提供一种基于MapReduce的色散并行实体提取系统及其方法,以通过基于MapReduce框架提取实体来保证缩短的实体提取响应时间。 构成:主服务器设备(100)通过将输入文档划分为目标文档数据来将目标文档数据分发到从服务器设备(200a-200N)。 从服务器设备将目标文档数据转换为能够在MapReduce框架中处理的数据格式,将转换的文档的内容分为句子,并将分割的句子划分为构造单元。 从服务器设备将构建单元的组合提取为实体候选,并定义提取的实体之间的关系。 (附图标记)(100)主服务器; (200a)从服务器1; (200b)从服务器2; (200N)从服务器N;

Patent Agency Ranking