Abstract:
PURPOSE: An extracting device for a semantic technological term of a high-capacity database is provided to enhance the effectiveness of a search by repetitively extracting technological term and complex-technological term and managing the extracted terms as technological intelligence. CONSTITUTION: An ARM(Acquisition Rule Manager) unit(100) searches a query language and outputs the searched query language. A TRS(Tech Retrieval System) unit(110) extracts the document group in which the query language is included and posting information. An ERA(External Resource Aggregator) unit(170) extracts the technological term, contextual information, associative relation, and technical report. The ERA unit provides the extracted information.
Abstract:
본 발명은 RDF 트리플 기반 확장 클래스-속성 관리 시스템 및 그 방법을 제공하기 위한 것으로, 온톨로지 스키마와 인스턴스들을 이용하여 클래스와 관계된 속성을 정의하는 클래스-속성 정의부와; 상기 클래스-속성 정의부에서 정의된 클래스-속성 관계를 대상으로 DBMS 에 적합한 저장 구조를 생성하고 관계 정보를 생성하여 저장하는 클래스-속성 적용부와; 상기 클래스-속성 적용부에서 생성된 관계 정보를 대상으로 질의 변환시 상기 클래스-속성 테이블을 참조하여 사용자 질의를 변환하고 질의를 처리하여 추론 서비스에 적합한 결과를 생성하고 제공하는 클래스-속성 활용부;를 포함하여 구성함으로서, RDF 트리플 저장구조로 인해 발생하는 비효율적인 정보의 접근을 개선하고, 온톨로지 스키마를 기반으로 추론 서비스에 적합하도록 클래스별 다중 속성을 생성하고, 생성된 클래스별 다중 속성 관계를 사용하여 질의를 처리함으로써 빠르고 효율적인 추론 서비스를 가능하게 한다. 온톨로지, 확장 클래스-속성, 추론 서비스, RDF 트리플, OWL 파서
Abstract:
본 발명은 다중 분류체계를 이용한 통제어 및 분야명 자동 할당 시스템 및 그 방법을 제공하기 위한 것으로, 다중 분류체계를 반입, 반출, 등록, 편집, 관리하는 다중 분류체계 관리부와; 상기 다중 분류체계 관리부에서 다중 분류체계를 제공받아 각 분류체계와 매핑된 통제어를 관리하는 통제어 관리부와; 상기 통제어 관리부에서 통제어를 제공받아 문서로부터 추출된 색인어와 통제어를 매칭하는 용어 매칭부와; 상기 용어 매칭부에서 매칭 결과를 제공받아 상위 통제어 및 각 분류체계 별 상위 분야명을 제시하는 통제어 및 분야명 할당부;를 포함하여 구성함으로서, 다중 분류체계 기반으로 문서를 분류할 필요가 있는 전 분야에 걸쳐 안정적이고 점진적인 성능 향상을 보일 수 있는 통제어 및 분야명 자동 할당을 가능하게 한다. 자동 할당, 통제어 할당, 분야명 할당, 분류체계, 문서 분류
Abstract:
A system and a method for assigning controlled terms and categories by using a multi-classification system are provided to enable a classification service provider to use a desired classification system by extracting core concepts from a document with the controlled terms, matching the core concepts with a multi-classification system, and assigning a plurality of categories to one document. A multi-classification system manager(10) receives, outputs, edits, and manages a multi-classification system, and a controlled term manager(20) manages controlled terms mapped to each classification system by receiving the multi-classification system from the multi-classification system manager. A term matcher(30) matches the controlled term with an index extracted from the document by receiving the controlled terms from the controlled term manager. A controlled term/category assigner(40) presents an upper controlled term and an upper category of each classification system by receiving a matching result from the term matcher. The multi-classification system manager registers or receives the multi-classification system, edits the multi-classification system without colliding an ID of each multi-classification system, provides the multi-classification system in real-time, and manages output of the multi-classification system.
Abstract:
본 발명은 지식베이스 구축 방법에 있어서, 스키마 항목을 정의하고, 파일, 난이도 및 도달 목표 중 적어도 하나를 포함하는 자원 정보가 입력되면, 그 자원 정보에 대한 고유한 식별자를 생성하고, 상기 입력된 자원 정보를 분석하여 속성 정보를 추출하고, 상기 추출된 각 속성 정보를 상기 정의된 스키마 항목에 따라 연관 관계를 설정하여 저장하는 것으로서, 분야 전문가들이 가지고 있는 특정 자원에 대한 난이도 정보, 현재 분야의 하부 요소 분야들에 대한 연구 순서에 대한 지식 등을 직접 지식베이스에 표현할 수 있다. 지식베이스, 자원, 속성, RDF
Abstract:
PURPOSE: An automatic query input system using vocabulary network, a method thereof, a recording medium storing a computer program about the method are provided to display various kinds of terms related to the search query in the below the query input window such as the meaning, substitute term, related term, super ordinate term, subordinate term, brotherly term, and words including the search term in the front and back. CONSTITUTION: A web interface(10) gets the input of user query. The web interface offers the linkage word list corresponding to the inputted query language as described above. The web interface automatically inputs the linkage word for to searching in the linkage word list of the offered query to the query input window. The web interface automatically searches the content corresponding to the inputted query and outputs words related to the search query.
Abstract:
PURPOSE: An automatic query input system using title, author, abstract information, a method for offering search result, and a recording medium storing a computer program of the method are provided to offer wide variety of information to users by displaying author information and abstract information through a separate window from the title list. CONSTITUTION: A user puts the mouse cursor on the title of the list of subject or clicks the title. A web interface(10) offers author and abstract information of the corresponding document through the separate window. When the user double-clicks the title on the title list, the web interface automatically inputs the selected title as described above to the query input window. At the same time, the web interface automatically searches the content corresponding to the title.
Abstract:
PURPOSE: A multi-entity centric integrated search system and a method thereof are provided to improve the satisfaction with respect to a search result corresponding to a query in which various entities are mixed. CONSTITUTION: A multi-entity centric integrated search system comprises an entity information acquiring server(10) and an integrated search result providing server(20). The entity information acquiring server includes a query input unit(11), a query analysis unit(12), an entity searcher(13), an entity/entity type detection unit and an entity/entity type determination unit. The query input unit receives a query. The query analysis unit analyzes the query to detect combinations of entities which are classified by a delimiter. The entity searcher searches an entity database with respect to the detected entities and then returns the existence of the entity and entity type as a result.
Abstract:
A history-based instance registering system and a method thereof are provided to improve the efficiency of edition and registration by editing and registering an instance with continuity based on the history through the ontology having mutual association. An instance registration setup unit(100) reads the ontology and stores it in a memory. The instance registration setup unit provides an interface which can edit instances within each class which is provided through an ontology viewer where the stored ontology relation is visualized. An instance registration editing unit(200) edits, searches and registers the instances by using the provided interface, and then extracts/edits object names from external information as managing the edited instances based on the history.
Abstract:
A system and a method for managing association information integrally by constructing an association information database are provided to enable an academic information service agent to quickly collect/monitor technical information needed for a service at a low cost and construct the database with the collected information by enabling each association to transmit the technical information automatically. An association information manager(110) enables a system manager to search and modify association information by searching an association list and detail information. An association member manager(120) enables the system manager to register association members and a database construction manager, and enables the association members to search/inquire the detail information. A multimedia data manager(130) manages multimedia data of each association. An association journal manager(150) searches bibliographic data to construct a database with association journals and technical data of the registered association members. An information processing manager(160) enables the system manager to construct the database in a batch mode by using a data management result received from the multimedia data manager and the association journal manager, and the data received from each association and association member.