• Journal of the Korean Data and Information Science Society
• /
• v.25 no.1
• /
• pp.187-194
• /
• 2014

We consider a Bayesian nonignorable selection model to accommodate the selection bias. Markov chain Monte Carlo methods is known to be very useful to fit the nonignorable selection model. However, sensitivity to prior assumptions on parameters for selection mechanism is a potential problem. To quantify the sensitivity to prior assumption, the deviance information criterion and the conditional predictive ordinate are used to compare the goodness-of-fit under two different prior specifications. It turns out that the 'MLE' prior gives better fit than the 'uniform' prior in viewpoints of goodness-of-fit measures.

• Journal of Korean Society of Transportation
• /
• v.14 no.1
• /
• pp.43-50
• /
• 1996

This paper is concerned with assessing goodness of fit of gravity models. The Chi-square test, or one of its asymptotic equivalents, is usually recommended for the purpose. A difficulty that frequently arises, particularly when working with urban travel data, is that the expected number of trips for most origin-destination(O-D) pairs are small. In order to test goodness of fit of gravity model, a simple approach, which depends on the number of O-D pairs and certain trip totals being large, is proposed in this paper. In addition, derivation of variance of Chi-square ratio is proposed to test the confidence interval of Chi-square ratio and application of its results with simulated data set is made to verify the usefulness of the results.

• Journal of the Korean Data and Information Science Society
• /
• v.14 no.2
• /
• pp.303-311
• /
• 2003

We consider the problem of testing cell probabilities in sparse multinomial data. Aerts, et al.(2000) presented $T_1=\sum\limits_{i=1}^k(\hat{p}_i-p_i)^2$ as a test statistic with the local polynomial estimator $(\hat{p}_i$, and showed its asymptotic distribution. When there are cell probabilities with relatively much different sizes, the same contribution of the difference between the estimator and the hypothetical probability at each cell in their test statistic would not be proper to measure the total goodness-of-fit. We consider a Pearson type of goodness-of-fit test statistic, $T=\sum\limits_{i=1}^k(\hat{p}_i-p_i)^2/p_i$ instead, and show it follows an asymptotic normal distribution.

• Journal of the Korean Data and Information Science Society
• /
• v.7 no.1
• /
• pp.37-46
• /
• 1996

In this paper, we introduce the goodness of fit test procedure for lifetime distribution using step stress accelerated lifetime data. Using the nonpapametric estimate of acceleration factor, we prove the strong consistence of empirical distribution function under null hypothesis. The critical vailues of Kolmogorov-Smirnov, Anderson-Darling, Cramer-von Mises statistics are computed when the lifetime distibution is assumed to be exponential and Weibull. The power of test statistics are compared through Monte-Cairo simulation study.

• Journal of the Korean Data and Information Science Society
• /
• v.9 no.2
• /
• pp.189-202
• /
• 1998

The goat of this paper is to study on variance estimation - Rice variance estimation, Gasser, Sroka and Jennen-Steinmetz's varince estimation - in smoothing goodness-of-fit tests. The comparisons of powers on test statistics are conducted by the change of variance, the number of oscillations, the amplitude of the alternative sample distribution.

• International Journal of Reliability and Applications
• /
• v.9 no.2
• /
• pp.125-140
• /
• 2008

The new better than used failure rate (NBUFR), Abouammoh and Ahmed (1988), and new better than average failure rate (NBAFR) Loh (1984) classes of life distributions, have been considered in the literature as natural weakenings of NBU (NWU) property. The paper considers testing exponentiality against strictly NBUFR (NBAFR) alternatives, or their duals, based on goodness of fit approach that is possible in life testing problems and that it results in simpler procedures that are asymptotically equivalent or better than standard ones. They may also have superior finite sample behavior. The asymptotic normality are proved. Powers, Pitman asymptotic efficiency and critical points are computed. Dealing with censored data case also studied. Practical applications of our tests in the medical sciences are present.

• Communications for Statistical Applications and Methods
• /
• v.8 no.2
• /
• pp.357-366
• /
• 2001

The present paper develops a method to check the propriety of link functions for binary data. In order to parameterize a certain type of goodness of the link, a family of link functions indexed by a shape parameter is proposed. I first investigate the maximum likelihood estimation of the shape parameter as well as regression parameters and then derive their large sample behaviors of the estimators. A score test is considered to evaluate the goodness of the current link function. For illustration, I employ two families of power transformations, the modulus transformation by John and Draper (1980) and the extended power transformation by Yeo and Johnson (2000), which are appropriate to detect symmetric and asymmetric inadequacy of the selected link function. respectively.

• Communications for Statistical Applications and Methods
• /
• v.21 no.4
• /
• 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.

• Ocean and Polar Research
• /
• v.36 no.4
• /
• pp.505-514
• /
• 2014

The development of deep seabed manganese nodules has been carried out with the aim of commercial development in 2023. It is important to forecast the price of the four metals (copper, nickel, cobalt, and manganese) extracted from manganese nodules because price change is a criterion for investment decision. The main purpose of the study is to forecast the price of four metals using the ARIMA model and VAR model, and calculate the MAPE to compare a goodness-of-fit between the two models. The estimated results of the two models reveal statistical significance and are in keeping with economic theory. The results of MAPE for goodness-of-fit show that the VAR model is between 0.1 and 0.2, and the ARIMA model is between 0.4 and 0.6. That is, the VAR model is better than the ARIMA model in forecasting changes in the price of metals.

• Communications for Statistical Applications and Methods
• /
• v.10 no.3
• /
• pp.955-969
• /
• 2003

The classical Fourier analysis, which is the typical frequency domain approach, is used to detect periodic trends that are of the sinusoidal shape in time series data. In this article, using a sequence of periodic step functions, describes an adaptive Fourier series where the patterns may take general periodic shapes that include sinusoidal as a special case. The results, which extend both Fourier analysis and Walsh-Fourier analysis, are applies to investigate the shape of the periodic component. Through the real data, compare the goodness-of-fit of the model using two methods, the adaptive Fourier method which is proposed method in this paper and classical Fourier method.