• Communications for Statistical Applications and Methods
• /
• v.2 no.2
• /
• pp.1-12
• /
• 1995

Let F be a life distribution with finite mean $\mu$ Then F is said to be in new better then worse than used in expectation (NBWUE(p)) class if $\varphi(u) {\geq} u$ for $0 {\leq}u{\leq}t_0$ and ${\varphi}(u) {\leq} u$ for $t_0< u {\leq} 1$ where ${\varphi}(u)$ is the scaled total-time-on-test transform and $p=F(t_0)$. We propose a testing procedure for $H_0$ : F is exponential against $H_1$ : NBWUE(p), and is not expontial, (or $H_1\;'$ : F is NWBUE (p), and is not exponential) using randomly censored data. Our procedure assumes kmowledge of the proportion p of the population that fail at or before the change-point $\t_0$. Know ledge of $\t_0$ itself is not assumed. The asymptotic normality of the test statistic is established and a Monte Carlo experiment is performed to investigate the speed of convergence of the test statistic to normality. The power of our test is also studied.

• Communications for Statistical Applications and Methods
• /
• v.17 no.6
• /
• pp.829-843
• /
• 2010

To test for the serial dependence in time series of counts data, Jung and Tremayne (2003) evaluated the size and power of several tests under the class of INARMA models based on binomial thinning operations for Poisson marginal distributions. The overdispersion phenomenon(i.e., a variance greater than the expectation) is common in the real world. Overdispersed count data can be modeled by using alternative thinning operations such as random coefficient thinning, iterated thinning, and quasi-binomial thinning. Such thinning operations can lead to time series models of counts with negative binomial or generalized Poisson marginal distributions. This paper examines whether the test statistics used by Jung and Tremayne (2003) on serial dependence in time series of counts data are affected by overdispersion.

• Communications for Statistical Applications and Methods
• /
• v.29 no.3
• /
• pp.301-317
• /
• 2022

In this paper, we propose a change point detection procedure based on the modified information criterion in a generalized lambda distribution (GLD) model. Simulations are conducted to obtain empirical critical values of the proposed test statistic. We have also conducted simulations to evaluate the performance of the proposed methods comparing to the log-likelihood method in terms of power, coverage probability, and confidence sets. Our results indicate that, under various conditions, the proposed method modified information criterion (MIC) approach shows good finite sample properties. Furthermore, we propose a new goodness-of-fit testing procedure based on the energy distance to evaluate the asymptotic null distribution of our test statistic. Two real data applications are provided to illustrate the use of the proposed method.

• Communications for Statistical Applications and Methods
• /
• v.27 no.4
• /
• pp.397-412
• /
• 2020

In this paper, a notion of data depth is used to propose nonparametric multivariate two sample tests for difference between scale parameters. Data depth can be used to measure the centrality or outlying-ness of the multivariate data point relative to data cloud. A difference in the scale parameters indicates the difference in the depth values of a multivariate data point. By observing this fact on a depth vs depth plot (DD-plot), we propose nonparametric multivariate two sample tests for scale parameters of multivariate distributions. The p-values of these proposed tests are obtained by using Fisher's permutation approach. The power performance of these proposed tests has been reported for few symmetric and skewed multivariate distributions with the existing tests. Illustration with real-life data is also provided.

• Journal of the Korean Statistical Society
• /
• v.22 no.1
• /
• pp.113-124
• /
• 1993

In this paper we show that the results of Ahn and Cho (1992) can be applied to a more general class of seasonal models, especially models with autocorrelated errors. Employing the idea of the "two-step estimation" method, we provide test statistics which are easy to compute and have the same asymptotic properties as those in Ahn and Cho (1992) for seasonal unit roots. A numerical example is presented to illustrate the methods and concepts. The power of the test statistics for finite samples is examined through a Monte Carlo sampling experiment.xperiment.

• Journal of the Korean Statistical Society
• /
• v.27 no.2
• /
• pp.165-187
• /
• 1998

Shin and Sarkar (1993, 1994) studied the problem of testing for a unit root in an AR(p) signal observed with MA(q) noise when the MA parameters are known. In this paper we consider the case when the MA parameters are unknown and to be estimated. Test statistics are defined using unit root parameter estimates based on three different estimation methods of Hannan and Rissanen (1982), Kohn (1979) and Shin and Sarkar (1995). An AR(p) process contaminated by MA(q) noise is a .estricted ARMA model, for which Shin and Sarkar (1995) derived an easy-to-compute Newton- Raphson estimator The two-stage estimation p.ocedu.e of Hannan and Rissanen (1982) is used to compute initial parameter estimates in implementing the iterative estimation methods of both Shin and Sarkar (1995) and Kohn (1979). In a simulation study we compare the relative performance of these unit root tests with respect to both size and power for p=q=1.

• The Korean Journal of Applied Statistics
• /
• v.24 no.2
• /
• pp.383-391
• /
• 2011

This paper presents a modified entropy-based test of fit for the inverse Gaussian distribution. The test is based on the entropy difference of the unknown data-generating distribution and the inverse Gaussian distribution. The entropy difference estimator used as the test statistic is obtained by employing Vasicek's sample entropy as an entropy estimator for the data-generating distribution and the uniformly minimum variance unbiased estimator as an entropy estimator for the inverse Gaussian distribution. The critical values of the test statistic empirically determined are provided in a tabular form. Monte Carlo simulations are performed to compare the proposed test with the previous entropy-based test in terms of power.

• The Korean Journal of Applied Statistics
• /
• v.11 no.2
• /
• pp.257-268
• /
• 1998

Many epidemiological studies investigate whether an association exists between a binary risk factor X and a binary response variable Y. They analyse whether an observed association between X and Y persists when the level of another factor Z that might influence the association is controlled. This involves testing conditional independence of X and Y controlling for Z. The Mantel-Haenszel test is most widely used to test conditional independence for sparse tables. But if the association between X and Y varies along the levels of Z, Mantel-Haenszel test has a low power problem. In this study, we propose an alternative test procedure which overcomes the low power problem in that case. We find out the null distribution of the alternative test statistic and compare its performance with the Mantel-Haenszel test by simulation.

• The Korean Journal of Applied Statistics
• /
• v.6 no.2
• /
• pp.289-302
• /
• 1993

A goodness-of-fit test for the two-parameter exponential distribution, for use with the singly Type I and Type II right censored samples, is proposed. The test statistic is based on the $L_1$-norm of discrepancy between the cumulative distribution function and the empirical distribution function. To deal with the unknown parameters problem, the K- transformation is considered and modified to be applied to the censored samples. Rosenblatt's transformation is extended to the cases of Type I and Type II censored samples, in order to transform the censored samples into the complete ones. The critial values of the test statistic are obtained by Monte Carlo simulations for some finite sample sizes. The power studies are conducted to compare the proposed test with the Pettitt(1977) test for exponentiality with censored samples. It appears that the proposed test has relatively good power properties for moderate and large sample sizes.

• Proceedings of the IEEK Conference
• /
• 2002.07c
• /
• pp.1531-1534
• /
• 2002

Shape description and its corresponding matching algorithm is one of the main concerns in MPEG-7. In this paper, a new method is proposed for shape registration of 2D objects for MPEG-7 Shapes are recognized using the Hu statistical moments in frequency domain. The Hu moments are moment-based descriptors of planar shapes, which are invariant under general translation, rotational, scaling, and reflection transformation. The image is transformed into frequency domain using Fourier Transform. Annular and radial wedge distributions fur the power spectra are extracted. Different statistical features (Hu moments) are found f3r the power spectrum of each selected transformed individual feature. The Euclidean distance of the extracted moment descriptors of the features are found with respect to the shapes in the database. The minimum Euclidean distance is the candidate for the matched shape. The simulation results are performed on the test shapes of MPEG-7.