• 응용통계연구
• /
• 제17권1호
• /
• pp.35-47
• /
• 2004

본 논문에서는 Kim & Bickel(2003)에서 제안한 이변량 정규분포를 위한 검정통계량을 Fattorini(1986)의 방법을 이용하여 이변량 이상인 경우에도 실제적으로 사용가능 하도록 일반화하였다. Fattorini(1986)의 통계량은 Shapiro & Wilk(1965)의 일변량 정규분포를 위한 검정통계량을 다변량으로 확장한 것이다. 그리고 제안된 통계량은 Fat-torini(1986) 통계량의 근사통계량으로 생각할 수 있으며 표본의 크기가 클 때도 사용 가능하다. 또한 모의실험을 통하여 여러 가지 대립가설에서 기존의 통계량과의 검정력을 비교하였다.

• 응용통계연구
• /
• 제23권3호
• /
• pp.487-495
• /
• 2010

본 논문에서는 지수분포의 검정에 자주 쓰이는 Shapiro와 Wilk (1972) 통계량과 이의 단점을 보완한 Kim (2001a)의 통계량을 위치모수가 주어지고 척도모수가 미지인 지수분포에서의 전진 제 2종 중도절단자료에 적용하였다. 이를 위하여 각각의 통계량을 Stephens (1978)을 이용하여 위치모수가 주어진 경우의 검정통계량으로 수정하고, 자료를 정규화 간격(normalized spacings)을 이용하여 변환하는 방법을 사용하였다. 모의실험을 통하여 검정력을 비교한 결과 Shapiro-Wilk 통계량보다 Kim (2001a)의 통계량을 이용할 때 고려한 거의 모든 경우 더 우수한 검정력을 나타내었다.

• 응용통계연구
• /
• 제21권2호
• /
• pp.265-273
• /
• 2008

본 논문에서는 Kim (2001a)에서 제안한 지수분포에서의 수정된 Shapiro와 Wilk (1972) $W_E$-통계량을 중도절단자료에 적용하였다. 검정통계량은 Samanta와 Schwarz (1988)에서 $W_E$-통계량을 중도절단자료에 대해 수정한 것과 같은 방법으로 정규화 등간격(normalized spacings)을 이용하여 수정하였다. 그 결과 제안된 통계량은 귀무가설에서 중도절단이 없는 경 우와 같은 분포를 갖고 표본크기만 변하게 된다. 제안된 통계량의 검정력을 Samanta와 Schwarz (1988)의 통계량과 비교한 결과, 중도절단이 없는 경우와 마찬가지로 중도절단이 있는 경우에도 변동계수가 1보다 크거나 같은 대립가설에서 제안된 통계량은 더 좋은 검정력을 나타내었다.

• Communications for Statistical Applications and Methods
• /
• 제8권2호
• /
• pp.473-481
• /
• 2001

Shapiro and Wilk (1972) developed a test for exponentiality with origin and scale unknown. The procedure consists of comparing the generalized least squares estimate of scale with the estimate of scale given by the sample variance. However the test statistic is inconsistent. Kim(2001) proposed a modified Shapiro-Wilk's test statistic based on the ratio of tow asymptotically efficient estimates of scale. In this paper, we study the asymptotic behavior of the statistic using the approximation of the quantile process by a sequence of Brownian bridges and represent the limit null distribution as an integral of a Brownian bridge.

• Communications for Statistical Applications and Methods
• /
• 제9권3호
• /
• pp.629-637
• /
• 2002

Shapiro and Wilk(1972) developed a test for exponentiality with origin and scale unknown. The procedure consists of comparing the generalized least squares estimate of scale with the estimate of scale given by the sample variance. However the test based on the statistic is inconsistent Kim(2001a) proposed a modified Shapiro-Wilk's test statistic using the ratio of two asymptotically efficient estimators of scale. In this paper, we study the consistency of the proposed test.

• Communications for Statistical Applications and Methods
• /
• 제10권3호
• /
• pp.765-777
• /
• 2003

In this paper, we investigate some measures as tests of multivariate normality based on principal components. The idea was proposed by Srivastava and Hui(1987). They generalized Shapiro-Wilk statistic for multi variate cases. We show the null distributions of the statistics do not depend on the unknown parameters and mention the asymptotic null distributions. Also power performance of the tests are assessed in a Monte Carlo study.

• Journal of the Korean Data and Information Science Society
• /
• 제11권2호
• /
• pp.139-155
• /
• 2000

An exclusive simulation study is conducted in computing means for order statistics in standard normal variate. Monte Carlo moments are used in Shapiro-Francia W' statistic computation. Finally, quantiles for Shapiro-Francia W' are generated. The study shows that in computing means for order statistics in standard normal variate, complicated distributions and intensive numerical integrations can be avoided by using Monte Carlo simulation. Lack of accuracy is minimal and computation simplicity is noteworthy.

• Journal of the Korean Data and Information Science Society
• /
• 제12권1호
• /
• pp.103-125
• /
• 2001

The Box-Cox transformation is a well known family of power transformation that brings a set of data into agreement with the normality assumption of the residuals and hence the response variable of a postulated model in regression analysis. This paper proposes a new method for estimating the Box-Cox transformation using maximization of the Sample Entropy statistic which forces the data to get closer to normal as much as possible. A comparative study of the proposed procedure with the maximum likelihood procedure, the procedure via artificial regression estimation, and the recently introduced maximization of the Shapiro-Francia W' statistic procedure is given. In addition, we generate a table for the optimal spacings parameter in computing the Sample Entropy statistic.

• Communications for Statistical Applications and Methods
• /
• 제6권3호
• /
• pp.901-908
• /
• 1999

Using the Transformed Lorenz curve which is introduced by Cho et al.(1999) we propose the test statistic for testing of normality that is very important test in statistical analysis and compare the proposed test statistic with the Shapiro and Wilk's W test statistic in terms of the power of test through by Monte Carlo method.

• Communications for Statistical Applications and Methods
• /
• 제6권3호
• /
• pp.687-694
• /
• 1999

The goodness of fit statistics of normality plots are obtained using the Receiver Operating Characteristic(ROC) method. This work is intended to compare with Shapiro-Wilk W statistic. Wel will use and discuss an accuracy of the test and the best cut-off value which minimizes the sum of the type I and II error probabilities.