• Journal of the Korean Data and Information Science Society
• /
• 제23권4호
• /
• pp.657-663
• /
• 2012

많은 분야에서 수집된 자료 중 데이터의 범위에서 많이 벗어난 아주 작은 값이나 아주 큰 값인 이상치가 종종 존재한다. 이런 이상치의 발생원인은 자료의 입력과정에서의 오류 또는 응답 과정에서 응답자의 특이한 답변 때문이다. 만약 자료에서 이상치가 존재할 경우 자료의 요약값인 평균과 분산에 많은 영향을 미쳐서 잘못된 정보가 산출된다는 문제점이 있다. 따라서 연구자는 자료에서 이상치가 존재하는지를 주의깊게 살펴보아야 한다. 특히 사람을 대상으로 실시한 임상자료의 경우 자료의 비용측면에서나 결과의 일관성 측면에서 이상치의 판단은 더욱 중요한 문제이다. 따라서 본 논문에서는 이상치를 판단하는 방법인 Grubb 검정과 Cochran 검정을 이용하여 임상자료에서의 이상치를 판단하는 방법을 소개하고자 한다.

• 한국IT서비스학회지
• /
• 제15권1호
• /
• pp.167-178
• /
• 2016

Spectrum sensing in cognitive radio (CR) has a great role in order to utilize idle spectrum opportunistically, since it is responsible for making available dynamic spectrum access efficiently. In this research area, collaboration among multiple cognitive radio users has been proposed for the betterment of detection reliability. Even though cooperation among them improves the spectrum sensing performance, some falsely reporting malicious users may degrade the performance rigorously. In this article, we have studied the detection and nullifying the harmful effects of such malicious users by applying some well known outlier detection methods based on Grubb's test, Boxplot method and Dixon's test in cooperative spectrum sensing. Initially, the performance of each technique is compared and found that Boxplot method outperforms both Grubb's and Dixon's test for the case where multiple malicious users are present. Secondly, a new algorithm based on reputation and weight is developed to identify malicious users and cancel out their negative impact in final decision making. Simulation results demonstrate that the proposed scheme effectively identifies the malicious users and suppress their harmful effects at the fusion center to decide whether the spectrum is idle.

• 대한치과의사협회지
• /
• 제55권9호
• /
• pp.604-616
• /
• 2017

In clinical dental research, errors occur in spite of careful study design and conduct. Data cleaning procedures intend to identify and correct these errors or at least to minimize their influence on study. Outlier is the one of these errors. Outlier detection is the first step in data analysis process which has a serious effect in the field of dental research. Hence, this paper aims to introduce the methods to detect the outliers and to examine their influences in statistical data analysis.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2004년도 추계학술대회 논문집
• /
• pp.626-629
• /
• 2004

Recently, the self-healing technique is being investigated to repair the damaged polymeric composites by the use of microcapsules with the healing agent. This technique can obtains both the damage detection and the damage repair simultaneously over the converntional repairing techniques. In this study, the effects of the catalyst ratio to the healing agent and thermal characteristics to the mixtures of healing agent are investigated through single lap shear tests and DSC. The Healing agents such as DCPD, ENB, and their mixtures are considered and Grubb's catalyst is used as a catalst.