• Journal of the Korean Society for information Management
• /
• v.26 no.3
• /
• pp.25-43
• /
• 2009

The purpose of this study is to suggest how the classic thesaurus structure of terms and links can be mined and updated from Wikipedia encyclopedia, which is the best practice of collective intelligence. In a comparison with ASIS&T thesaurus, it was found that Wikipedia contains a substantial coverage of domain-specific concepts and semantic relations. Furthermore, it was resulted that the structural characteristics of Wikipedia, such as redirects, categories, and mutual links are suitable to extract semantic relationships of thesaurus. It is needed to apply to update various thesauri, including multilingual thesaurus, in order to generalize the results of this research.

• Journal of the Daesoon Academy of Sciences
• /
• v.19
• /
• pp.21-45
• /
• 2005

The purpose of this study is to examine the establishment and development of Daesoon Thought Thesaurus. Specifically, this study examined the matters to be considered in the stage of Thesauri planning according to the Thesauri Construction process : presents the methods and standards of Thesauri Construction according to processes such as identification of the indexing policy, establishment of Thesauri system, collection of vocabulary, selection of preferred term, clustering of the terms, establishment of term relationships, overall adjustment, Thesauri test, proofreading by professional display, maintenance and updating. Since religion information is unique or totally different from the information in other areas, it is most important to construct the Thesauri suitable for system after carefully recognizing the concept of religion terms.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2004.05a
• /
• pp.321-324
• /
• 2004

본 연구는 재사용이 가능한 객체지향 컴포넌트의 효율적인 검색을 위해 클래스 개념 범주(CCC)를 정의하고 클래스의 상속 관계를 이용한 CCC 상속을 제안하였다. CCC의 상속은 클래스 간 상속관계를 인지한 후, 하위 클래스는 상위 클래스의 모든 CCC를 자동 상속받게 되며, 클래스 구문분석에 의해 객체지향 코드로부터 자동 인식된다. 본 연구는 모든 용어에 대한 범주를 수동으로 할당해주는 기존 방법의 단점을 극복할 수 있으며, 시소러스를 자동으로 갱신할 수 있다는 장점이 있다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.10d
• /
• pp.673-675
• /
• 2002

본 논문에서는 전자메일을 사용자 적합도(선호도)를 기준으로 분류함에 있어 좀더 사용자 선호도를 반영할 수 있는 시스템 구조를 제안한다. 사용자 선호도는 2단계에 걸쳐서 반영되는데, 1단계에서는 사용자 관련메일로 판단된 메일정보추출어구(MIWs)들로부터 사용자 동적 시소러스(DS)의 갱신을 통해 이뤄지며, 2단계에서는 DS로부터 추출된 키워드들을 갖고 유전자 알고리즘을 작동시킬 때, 사용자선호도 feedback을 받음으로서 이뤄진다. 테스트는 kaist뉴스그룹으로부터 임의로 추출된 5개 분야 10개씩의 메일을 sample로 사용하였으며, DS로부터 추출된 키워드가 유전자알고리즘 모듈을 통해 사용자 feedback을 받았을 때, 세대가 거듭함에 따라 사용자가 요구하는 threshold 값에 근사하게 관련키워드들이 수집되었다. 그 결과 사용자 전자메일분류시스템(PECS)의 성능도 폴더정보키워드(FIWs)의 변화에 따라 향상될 수 있음을 확인하였다.

• Journal of the Korean BIBLIA Society for library and Information Science
• /
• v.9 no.1
• /
• pp.221-240
• /
• 1998

This study concerns information quality of the database which has been produced in the ASKERIC database. The measures used in this study were accuracy of the records. consistency, completeness and currency. Accuracy was measurred in terms of the keywords used in different ways in the US and Britain and the spelling errors in the records. Consistency was measured in terms of 'see also' and 'see reference' mechanism and character capitalization. Completeness was measured as follows： completeness of the search fields in the record and relevance of search fields. Currency was measured using the publication date. The experimental result showed that ERIC databases had some errors in terms of accuracy. consistency. completeness and currency. However, continuous striving for the automatic error checking functions and the policy of feedback from users have contributed to the improvement of the quality in ERIC databases.