http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
이도헌(Lee Do Heon),조일래(Cho Il Rae),김종덕(Kim Jong Deok) 한국정보처리학회 1997 정보처리학회논문지 Vol.4 No.4
The mining of association rules discovers the tendency of events occurring simultaneously in large databases. Previously announced research on association rules deals with associations with respect to the whole transaction. However, some association rules could have very high confidence in a sub-range of the time domain, even though they do not have quite high confidence in the whole time domain. Such kind of association rules are expected to be very useful in various decision making problems. In this paper, we define transient association rule, as an association with high confidence worthy of special attention in a partial time interval, and propose an efficient algorithm which finds out the time intervals appropriate to transient association rules from large databases. We propose the data-driven retrieval method excluding unnecessary interval search, and design an effective data structure manageable in amin memory obtained by one scanning of database, which offers the necessary information to next retrieval phase. In addition, our simulation shows that the suggested algorithm has reliable performance at the time cost acceptable in application areas.
퍼지 개념 계층을 기반으로 한 데이타베이스 속성 상호 관계의 발견
이도헌(Do Heon LEE),김명호(Myoung Ho KIM) 한국정보과학회 1995 정보과학회논문지 Vol.22 No.4
본 논문은 데이타베이스로부터 속성간의 의미적인 상호관계를 발견하는 방법을 제안한다. 속성간의 상호관계는 퍼지용어로 기술되는 서술규칙(descriptive rule)의 형태로 정형화된다. 타당한 서술규칙을 판별하기 위한 척도로서 지지도(support degree)와 확신도(confidence factor)가 제안된다. 아울러, 서술규칙의 특성상 발생할 수 있는 과대해석의 문제를 해결하기 위하여 범위도(coverage)라는 척도가 정의된다. 삼단계로 이루어진 프로시듀어를 통해 서술규칙을 발견할 수 있는데, 첫번째 단계는 퍼지가설정제법(fuzzy hypothesis refinement algorithm)을 적용하여, 인증된 퍼지가설(qualified fuzzy hypothesis)을 발견하는 작업이다. 두번째 단계는 인증된 퍼지가설로부터 “if 전제부 then 결과부”로 표현되는 서술규칙을 도출하는 작업이고 마지막 단계는 도출된 결과들을 발견목적에 따라 설명규칙(explanatory rule), 분류규칙(classification rule), 요약(data summarization)과 같이 해석하는 작업이다. 본 연구의 주안점은 좀더 이해하기 쉬운 발견결과의 획득과 실제 현장지식(domain knowledge)의 더욱 효과적인 활용에 있다. This paper proposes a discovery mechanism for semantic relationships among several attributes of databases. Attribute relationships are formalized in forms of descriptive rules with fuzzy terms. We provide two measures called support degree and confidence factor to identify valid descriptive rules. In addition, the notion of coverage is introduced to reduce the possibility of exaggerated interpretation. A canonical three-stage procedure for discovering descriptive rules are presented. The first stage is to discover qualified fuzzy hypotheses and the second is to elicit descriptive rules from such qualified hypotheses. The final stage is to interpret the results as explanatory rules, classification rules and data summarization according to the discovery purpose. Our main concern is to achieve comprehensible representation of discovered rules and effective utilization of real domain knowledge.