• Communications for Statistical Applications and Methods
• /
• v.6 no.2
• /
• pp.523-532
• /
• 1999

Generalized linear models have various applications for data arising from many kinds of statistical studies. Although the response variable is generally assumed to be generated from a wide class of probability distributions we focus on count data that are most often analyzed using binomial models for proportions or poisson models for rates. The methods and results presented here also apply to many other categorical data models in general due to the relationship between multinomial and poisson sampling. The novelty of the approach suggested here is that all conditional distribution s can be specified directly so that staraightforward Gibbs sampling is possible. The prior distribution consists of two stages. We rely on a normal nonconjugate prior at the first stage and a vague prior for hyperparameters at the second stage. The methods are demonstrated with an illustrative example using data collected by Rosenkranz and raftery(1994) concerning the number of hospital admissions due to back pain in Washington state.

• Journal of Korean Society for Quality Management
• /
• v.26 no.2
• /
• pp.39-50
• /
• 1998

A screening procedure, where one or more correlated variables are used for screeing, is reviewed from the point of statistical hypothesis testing. Without assuming a specific probability model for the joint distribution of the performance and screening variables, some principles are provided to establish the best screeing region. A, pp.ication examples are provided for two cases; ⅰ) the case where the performance variable is dichotomous and ⅱ) the case where the performance variable is continuous. In case ⅰ), a normal model is assumed for the conditional distribution of the screening variable given the performance variable. In case ⅱ), the performance and screening variables are assumed to be jointly normally distributed.

• The Korean Journal of Applied Statistics
• /
• v.15 no.1
• /
• pp.187-199
• /
• 2002

When the sample size is small, exact tests are often employed because the asymptotic distribution of the test statistic is in doubt. The advantage of exact tests is that it is guaranteed to bound the type I error probability to the nominal level. In this paper we review the methods of constructing exact tests, the algorithm and commercial software. We also examine the difference between exact p-values obtained from exact tests and true p-values obtained from the true underlying distribution.

• Journal of Applied Reliability
• /
• v.8 no.2
• /
• pp.75-85
• /
• 2008

In this paper, we compared domestic with foreign pilot lamps installed on the instrument board or electronic modules (car audio, air-conditional system, etc.) of an automotive vehicle by an accelerated life test in order to estimate the life of domestic pilot lamps. An accelerated life test method was developed and the relation of the life and voltage stress was analyzed. The main results are as follows; i) $B_{10}$ life of pilot lamp is above 5,000 hours, vehicle travel time for 10 years. ii) the life of domestic pilot lamp is longer than that of foreign thing. iii) the life distribution of domestic pilot lamp is wider than that of foreign thing. iv) it is possible to promote import replacement of automobile pilot lamp.

• Communications for Statistical Applications and Methods
• /
• v.13 no.2
• /
• pp.255-266
• /
• 2006

This paper proposes a class of distributions which is useful in making inferences about the sum of values from a normal and a doubly truncated normal distribution. It is seen that the class is associated with the conditional distributions of truncated bivariate normal. The salient features of the class are mathematical tractability and strict inclusion of the normal and the skew-normal laws. Further it includes a shape parameter, to some extent, controls the index of skewness so that the class of distributions will prove useful in other contexts. Necessary theories involved in deriving the class of distributions are provided and some properties of the class are also studied.

• Journal of Communications and Networks
• /
• v.18 no.2
• /
• pp.246-260
• /
• 2016

A major challenge achieving scalable access control for a large number of subscribers in a public broadcast is to distribute key update messages reliably to all stateless receivers. However, in a public broadcast, the rekeying messages can be dropped or compromised during transmission over an insecure broadcast channel, or transmitted to receivers while they were off-line. In this study, we propose a novel group key management scheme. It features a mechanism to allow legitimate receivers to recover the current group key, even if they lose key update messages for long-term sessions. The scheme uses short hint messages and member computation. Performance analysis shows that the proposed scheme has the advantages of scalability and efficient rekeying compared to previous reliable group key distribution schemes. The proposed key management scheme targets a conditional access system in a media broadcast in which there is no feedback channel from receivers to the broadcasting station.

• Communications for Statistical Applications and Methods
• /
• v.30 no.6
• /
• pp.605-629
• /
• 2023

This paper proposes some diagnostics procedures for the skew-t linear regression model with censored response. The skew-t distribution is an attractive family of asymmetrical heavy-tailed densities that includes the normal, skew-normal and student's-t distributions as special cases. Inspired by the power and wide applicability of the EM-type algorithm, local and global influence analysis, based on the conditional expectation of the complete-data log-likelihood function are developed, following Zhu and Lee's approach. For the local influence analysis, four specific perturbation schemes are discussed. Two real data sets, from education and economics, which are right and left censoring, respectively, are analyzed in order to illustrate the usefulness of the proposed methodology.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2019.05a
• /
• pp.151-151
• /
• 2019

Because drought is a complex and stochastic phenomenon in nature, statistical approaches for drought assessment receive great attention for water resource planning and management. Generally drought characteristics such as severity, duration and intensity are modelled separately. This study aims to develop a relationship between drought characteristics using a bivariate copula model. To achieve the objective, we calculated the Standardized Precipitation Index (SPI) using rainfall data at 6 rain gauge stations for the period of 1961-1999 in Jehlum River Basin, Pakistan, and investigated the drought characteristics. Since there is a significant correlation between drought severity and duration, they are usually modeled using different marginal distributions and joint distribution function. Using exponential distribution for drought severity and log-logistic distribution for drought duration, the Galambos copula was recognized as best copula to model joint distribution of drought severity and duration based on the KS-statistic. Various return periods of drought were calculated to identify time interval of repeated drought events. The result of this study can provide useful information for effective water resource management and shows superiority against univariate drought analysis.

• Journal of Digital Convergence
• /
• v.11 no.11
• /
• pp.409-414
• /
• 2013

Naive Bayesian classifiers are a powerful and well-known type of classifiers that can be easily induced from a dataset of sample cases. However, the strong conditional independence assumptions can sometimes lead to weak classification performance. Normally, naive Bayesian classifiers use Gaussian distributions to handle continuous attributes and to represent the likelihood of the features conditioned on the classes. The probability density of attributes, however, is not always well fitted by a Gaussian distribution. Another eminent type of classifier is the neuro-fuzzy classifier, which can learn fuzzy rules and fuzzy sets using supervised learning. Since there are specific structural similarities between a neuro-fuzzy classifier and a naive Bayesian classifier, the purpose of this study is to apply learning distribution graphs constructed by a neuro-fuzzy network to naive Bayesian classifiers. We compare the Gaussian distribution graphs with the fuzzy distribution graphs for the naive Bayesian classifier. We applied these two types of distribution graphs to classify leukemia and colon DNA microarray data sets. The results demonstrate that a naive Bayesian classifier with fuzzy distribution graphs is more reliable than that with Gaussian distribution graphs.

• Journal of Distribution Science
• /
• v.12 no.4
• /
• pp.73-83
• /
• 2014

Purpose - This study aims to classify and extend the consumer value of movie theaters into various values such as functional value, emotional value, social value, epistemic value, and conditional value based on the theory of consumption value by Sheth, Newman and Gross (1991). It also aims to verify the path structure of consumption value→customer satisfaction→behavior intention of movie theaters to confirm its generalization. Research design, data, and methodology - This study was conducted by collecting data on Kimpo university students from various areas in Incheon, Northern Seoul, Ilsan, Kyonggi Province, and Kimpo City. The survey was conducted by distributing 280 survey papers from Oct. 5 to 15, 2013 and collecting 238 of them. The final analysis used 208 questionnaires, after excluding 30 invalid responses. The statistical analysis of this study used the SPSS 19.0 statistics package. Results - The results of the survey are as follows: First, consumption values of movie theaters are classified into the following five groups: functional value, emotional value, social value, epistemic value, and conditional value. This study verified that consumption values play a role as a previous variable of customer satisfaction. Second, functional value, emotional value, and epistemic value have positive effects on customer satisfaction. On the other hand, social value and conditional value do not affect customer satisfaction. Finally, customer satisfaction has a positive impact on behavior intention. Theater users have an intention to re-use or recommend the movie theater they used when they are satisfied with a movie theater's physical environment and services. Conclusions - This study can provide academic and practical implications as follows based on the results mentioned above. First, academic implications can be found in that consumption values of movie theater users are classified into five values based on the theory of consumption value by Sheth et al. (1991). In the previous study, the service quality of a movie theater was studied based on the service quality of service encounters and a physical environment→customer satisfaction→behavior intention path structure. However, this study was verified by a consumption value→customer satisfaction→behavior intention path structure to classify consumption value, but not service quality or perceived value of quality, to confirm this generalization. Second, practical implications can be found in that the relative impact of consumption value of movie theaters on consumer satisfaction showed that functional value was followed by epistemic value and emotional value. In the previous study on movie theaters, previous variables of customer satisfaction were separated only by functional service quality including service encounters and physical environment; in some other studies, quality of service encounter had a direct effect on customer satisfaction. Accordingly, a marketing manager of a movie theater should develop various differentiated services by reflecting not only functional value such as service encounters and physical environment but also epistemic value and emotional value.