• Communications for Statistical Applications and Methods
• /
• 제15권4호
• /
• pp.531-542
• /
• 2008

시계열 자료를 분석할 때, 시계열 자료가 가지고 있는 추세를 제거하기 위하여 결정적 추세인 경우 회귀모형을 이용하고, 확률적 추세인 경우 차분하는 방법을 이용한다. 이 때 제거의 옳바른 기준이 되는 검정 방법이 단위근 검정이다. 그러나 기존의 Dickey-Fuller 검정 (Dickey와 Fuller, 1979)은 표본 수가 작고, 단위근에 가까울 경우 검정력이 낮으며, 베이지안 단위근 검정은 절차가 복잡하다. 본 논문에서는 기존의 두 방법들의 문제점을 해결하기 위하여, 전통적 Dickey-Fuller 검정 방법과 베이지안 방법을 결합한 형태의 검정방법으로 제안하였다. 제안된 검정방법은 모형 AR(1)에서 계수가 거의 1이거나 표본 수가 작을 경우, 기존의 Dickey-Fuller 검정보다 검정력이 높을 뿐만 아니라 일반적인 베이지안 방법 보다 절차가 간단한 검정법이 된다.

• Communications for Statistical Applications and Methods
• /
• 제5권3호
• /
• pp.707-722
• /
• 1998

This paper evaluates a new clinical trial designs for low infection rate disease. This type of sparse disease reaction makes the traditional two sample t-test or Wilcoxon rank-sum test inefficient compared to a new test suggested. The new test, which is based solely on the larger changes, is shown to be more effective than existing method by simulation for small samples. However, this test can be shown to be connected to the locally most powerful rank test under certain practical conditions. This design is motivated in testing the treatment effects in periodontal disease research.

• Communications for Statistical Applications and Methods
• /
• 제23권3호
• /
• pp.215-230
• /
• 2016

The powers of some tests for independence hypothesis against positive (negative) quadrant dependence in generalized Farlie-Gumbel-Morgenstern distribution are compared graphically by simulation. Some of these tests are usual linear rank tests of independence. Two other possible rank tests of independence are locally most powerful rank test and a powerful nonparametric test based on the $Cram{\acute{e}}r-von$ Mises statistic. We also evaluate the empirical power of the class of distribution-free tests proposed by Kochar and Gupta (1987) based on the asymptotic distribution of a U-statistic and the test statistic proposed by $G{\ddot{u}}ven$ and Kotz (2008) in generalized Farlie-Gumbel-Morgenstern distribution. Tests of independence are also compared for sample sizes n = 20, 30, 50, empirically. Finally, we apply two examples to illustrate the results.

• 대한기계학회논문집A
• /
• 제27권7호
• /
• pp.1200-1209
• /
• 2003

In this paper, a unified methodology based on the local stress concept to estimate residual strength of locally thinned pipes. An underlying idea of the proposed methodology is that the local stress in the minimum section for locally thinned pipe is related to the reference stress, popularly used in creep problems. Then the problem remains how to define the reference stress, that is the reference load. Extensive three-dimensional finite element (FE) analyses were performed to simulate full-scale pipe tests conducted for various shapes of wall thinned area under internal pressure and bending moment. Based on these FE results, the reference load is proposed, which is independent of materials. A natural outcome of this method is the maximum load capacity. By comparing with existing test results, it is shown that the reference stress is related to the fracture stress, which in turn can be posed as the fracture criterion of locally thinned pipes. The proposed method is powerful as it can be easily generalised to more complex problems, such as pipe bends and tee-joints.

• Journal of the Korean Statistical Society
• /
• 제33권2호
• /
• pp.149-157
• /
• 2004

The stationarity is one of the most important properties of a time series. We propose robust sign tests for seasonal autoregressive processes to determine whether or not a time series is stationary. The proposed tests are robust to the outliers and the heteroscedastic errors, and they have an exact binomial null distribution regardless of the period of seasonality and types of median adjustments. A Monte-Carlo simulation shows that the sign test is locally more powerful than the tests based on ordinary least squares estimator (OLSE) for heavy-tailed and/or heteroscedastic error distributions.

• 한국통계학회:학술대회논문집
• /
• 한국통계학회 2003년도 춘계 학술발표회 논문집
• /
• pp.281-286
• /
• 2003

The stationarity is one of the most important properties of a time series. We propose robust sign tests for seasonal autoregressive process to determine whether or not a time series is stationary. The tests have an exact binomial null distribution and are robust to the outliers and the heteroscedastic errors. Monte-Carlo simulation shows that the sign test is locally more powerful than the OLSE-based tests for heavy-tailed and/or heteroscedastic error distributions.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제30권7_8호
• /
• pp.773-779
• /
• 2003

얼굴 영역의 정확한 위치를 정확히 찾는 것은 얼굴 인식을 위한 핵심적인 전처리 과정이다. 본 논문에서는 조명조건, 표정, 배경의 변화에 무관하게 얼굴의 구성요소를 검출할 수 있는 강건한 기법을 제안한다. 수평 방향의 상대적으로 어두운 화소값을 갖는 눈썹, 눈, 입 둥과 같은 독립된 얼굴 요소를 검출하기 위해 형상 분해 국부 최적 임계치 기법을 적용하며 얼굴을 구성하는 간단한 기하학적 조건을 만족하는 얼굴 요소의 그룹을 검색함으로써 순수 얼굴영역을 검출한다. AR-face 데이타베이스의 영상에 적용한 결과, 두꺼운 안경테에 의해 눈썹이 가리워진 특수한 경우를 제외한 거의 모든 영상에 대해 정확한 얼굴 영역을 검출할 수 있었고, 얼굴의 대칭성을 이용해 회전과 이동 변화를 보상할 수 있으므로 강건한 얼굴 인식의 전처리 과정으로 사용할 수 있다.