• Communications for Statistical Applications and Methods
• /
• 제19권4호
• /
• pp.607-617
• /
• 2012

The goodness-of-fit test for multivariate normal distribution is important because most multivariate statistical methods are based on the assumption of multivariate normality. We propose goodness-of-fit test statistics for multivariate normality based on the modified squared distance. The empirical percentage points of the null distribution of the proposed statistics are presented via numerical simulations. We compare performance of several test statistics through a Monte Carlo simulation.

• 한국통계학회:학술대회논문집
• /
• 한국통계학회 2003년도 추계 학술발표회 논문집
• /
• pp.233-238
• /
• 2003

Goodness of fit test statistics based on the information discrepancy have been shown to perform very well (Vasicek 1976, Dudewicz and van der Meulen 1981, Chandra et al 1982, Gohkale 1983, Arizona and Ohta 1989, Ebrahimi et al 1992, etc). Although the test is well defined for the non-censored case, censored case has not been discussed in the literature. Therefore we consider a goodness of fit test based on the partial Kullback-Leibler(KL) information with the type II censored data. We derive the partial KL information of the null distribution function and a nonparametric distribution function, and establish a goodness of fit test statistic. We consider the exponential and normal distributions and made Monte Calro simulations to compare the test statistics with some existing tests.

• 한국산업보건학회지
• /
• 제9권2호
• /
• pp.79-89
• /
• 1999

A quantitative fit test, condensation nuclei counting (Portacount 8025, TSI), was performed concurrently with a banana oil (isoamyl acetate: IAA) qualitative fit test (MSA) to evaluate reliability on IAA QLFT and correlation between two methods. One brands of quarter mask (3M model 7500 medium) was prepared for QLFT with HEPA filter and gas & vapor removing media, i.e., combination cartridge. 110 subjects (65 male, 45 female) were fit tested QNFT and QLFT each three times. For a wearer combination having a FF<10, as determined by CNC QNFT, the point es timate (${\beta}$-error) of the probability of that combination not being rejected by the banana oil QLFT was found to be 0.0 with 95% confidence that this statistic is not expected to exceed 0.15. For a wearer combination having a FF<100, as determined by CNC QNFT, the point estimate of the probability of that combination not being rejected by the banana oil QLFT was found to be 0.07 with 95% confidence that this statistic is not expected to exceed 0.13. The uncertainty associated with each estimate, however, is large due to the small number of study subjects with inadequately fitting respirators.

• Safety and Health at Work
• /
• 제13권4호
• /
• pp.500-506
• /
• 2022

This study evaluated the consistency between two quantitative fit test devices with different methods of ambient aerosol counting. Three types of respirators (N95, half mask, and full facepiece) were worn by 50 participants (male, n = 25; female, n = 25), PortaCount (Pro+ 8038) and MT (05U) were connected to one probe to one mask, and fit factors (FFs) were measured simultaneously with the original and modified protocols. As a result of comparing MT FFs with PortaCount FFs as references and by applying for the pass/fail criteria (FF = 100), the consistency between the two devices for half masks and full facepieces was very high. N95 was somewhat weaker than the two type of respirators in the consistency; however, the correlation between the two devices was very strong (p < 0.0001). The results showed that an FF of 100 as measured by PortaCount was likely to be measured as 75 by the MT. Therefore, when performing the fit test for N95 using the MT and pass level of FF 100, a certain level of adjustment is necessary, whether end-user or putting a scaling factor by manufacturer.

• 물과 미래
• /
• 제7권2호
• /
• pp.107-111
• /
• 1974

Annual runoff in the Nakdong river basin has been analyzed to find the probability functions of best fit to distribution of historical annual runoff. The results obtained are as follows; (1) Log-normal 3-parameter disrtibution is believed as the probability function of best fit to historical distribution (2) Log-normal 3-parameter disrtibution is believed as the best fit probability function among Log-normal dist-ributions. (3) In the test of goodness of fit, $x^2-test$ shows that probability of $x^2-valus$ in Log-normal 3-parameter distribution is nearly more than 90%. But in the Simirnov-Kolmogorov test, hypotheses for the probability distributions cannot be rejected at significance level 5% & 1%. (4) Among 7 gauging stations, Dongchon & Koryung-Bridge's records show lower fitness to the theoretical probability functions than other 5 gauging station's

• Journal of Preventive Medicine and Public Health
• /
• 제31권3호
• /
• pp.440-448
• /
• 1998

In many developed countries, for example, USA, respirator fit testing is required before entering specific work environment to ensure that the respirator worn satisfies a minimum of fit and that the user knows when the respirator fits properly. Unfortunately because we have not fit test regulation in Korea, a lot of workers wearing respirators may be potentially exposed to hazards. This study was conducted to evaluate the fitting performance for respirators and correlation fit factors with facial dimensions of wearers. 110 subjects (70 males, 40 females) were fit tested for three quarter masks, i.e., two domestic-made Mask 2, and Y and one foreign-made Mask T using PortaCount 8020. A facial dimension survey of the same subjects was conducted to develop a facial dimension grids fer correlation fit factors with facial dimension parameters. A facial dimension grid was developed on the basis of face length and lip length for quarter masks. The results obtained were as follows : 1 Fit factors of Mask T were much higher than those of Masks Z, and Y. 2. Males were fitted more properly than females. 3. Male in box 'f' of grid would be adequately fitted Mask Y and male in box 'b', 'e', 'f', 'h' of grid would be sufficiently fitted Mask T. Female in box 'h' of grid may have a good fitting performance for both Mask Y, and T. But subjects in all boxes of grid would be inadequately fitted Mask Z.

• 한국의류학회지
• /
• 제28권5호
• /
• pp.637-647
• /
• 2004

The goal of this study was to develop bicycle apparel pattern for women cyclists. To investigate the pattern of bicycle apparel in Korea, anthropometric data were collected using direct and indirect measurements from October 2002 to July 2003. Three female subjects were recruited from forty measured women cyclists for pilot test. The differences between the responses of prototype bicycle apparel and existing bicycle apparel were compared using mean, standard deviation and t-test. The outfits for cyclists was completed through the three times rectification. Based on the results from the fit and appearance testing, I concluded as follows: The mean value of the center back length, back width, sleeve out-seam, bust ＆ thigh circumference, and weight of cyclists are more than the mean value of the non-cyclists. The result of fit evaluation by judges shows that there is a significant differences in the response to the shoulder, collar, movement of arms of jersey and movement of upper legs, crotch fit and general fit of shots. The result of appearance test between two outfits is no differences the mean value except the waist ease when sitting on cycle. Prototype bicycle apparel pattern for women cyclists developed and tested in this study have the potential to increase field performance and appearance.

• Journal of the Korean Data and Information Science Society
• /
• 제28권4호
• /
• pp.733-742
• /
• 2017

통계학에서 사용되어지고 있는 Gumbel 분포는 환경과학, 시스템 신뢰성, 수문학과 같은 분야에서 많이 응용되고 있다. 따라서 환경과학, 시스템 신뢰성, 수문학과 관련된 자료를 분석함에 있어서 분석에 사용되어지는 자료가 Gumbel 분포를 따르는지 확인하는 것은 매우 중요하다. 이를 확인하기 위해 본 논문에서는 새로운 두 가지의 Gumbel 분포의 적합도 검정통계량을 일반화된 로렌츠 곡선을 기반으로 하여 제안하였고, Anderson - Darling 검정, Cramer - vonMises 검정, 수정된 Anderson - Darling 검정과 비교하였다. 그 결과 새롭게 제안한 검정통계량은 기존의 검정방법에 비하여 우수한 것을 확인할 수 있었다. 또한 새롭게 제안한 변형된 표본 일반화된 로렌츠 곡선을 이용하여 두 가지의 새로운 적합도 검정 그래프 방법을 제안하였고, 새롭게 제안된 그래프를 통하여 손쉽게 데이터가 Gumbel 분포를 따르는지를 파악 할 수 있었다. 또한 호주 시드니의 연간 일 최대 강수량 자료를 사용하여 새롭게 제안한 검정 통계량과 그래프 방법을 이용하여 적용해 보았다.

• Journal of the Korean Data and Information Science Society
• /
• 제13권1호
• /
• pp.113-119
• /
• 2002

A powerful and easily computed goodness-of-fit test for Pareto distribution which does not depend on the unknown location and scale parameters is proposed based on the transformed sample Lorenz curve. We compare the power of the proposed test statistic with the other goodness-of-fit tests for Pareto distribution against various alternatives through Monte Carlo methods.

• Communications for Statistical Applications and Methods
• /
• 제21권4호
• /
• pp.309-316
• /
• 2014

Testing normality is very important because the most common assumption is normality in statistical analysis. We propose a new plot and test statistic to goodness-of-fit test for normality based on the generalized Lorenz curve. We compare the new plot with the Q-Q plot. We also compare the new test statistic with the Kolmogorov-Smirnov (KS), Cramer-von Mises (CVM), Anderson-Darling (AD), Shapiro-Francia (SF), and Shapiro-Wilks (W) test statistic in terms of the power of the test through by Monte Carlo method. As a result, new plot is clearly classified normality and non-normality than Q-Q plot; in addition, the new test statistic is more powerful than the other test statistics for asymmetrical distribution. We check the proposed test statistic and plot using Hodgkin's disease data.