We propose a semi-parametric model to analyze clustered and interval-censored data; in addition, we plugged-in a gamma frailty to the model to measure the association of members within the same cluster. We propose an estimation procedure based on EM algorithm. Simulation results showed that our estimation procedure may result in unbiased estimates. The standard error is smaller than expected and provides conservative results to estimate the coverage rate; however, this trend gradually disappeared as the number of members in the same cluster increased. In addition, our proposed method was illustrated with data taken from diabetic retinopathy studies to evaluate the effectiveness of laser photocoagulation in delaying or preventing the onset of blindness in individuals with diabetic retinopathy.

Agreement analysis is conducted to assess reliability among rating results performed repeatedly on the same subjects by one or more raters. The kappa statistic is commonly used when rating scales are categorical. The simple and weighted kappa statistics are used to measure the degree of agreement between two raters, and the generalized kappa statistics to measure the degree of agreement among more than two raters. In this paper, we compare the performance of four different generalized kappa statistics proposed by Fleiss (1971), Conger (1980), Randolph (2005), and Gwet (2008a). We also examine how sensitive each of four generalized kappa statistics can be to the marginal probability distribution as to whether marginal balancedness and/or homogeneity hold or not. The performance of the four methods is compared in terms of the relative bias and coverage rate through simulation studies in various scenarios with different numbers of raters, subjects, and categories. A real data example is also presented to illustrate the four methods.

The successful launch of Arirang satellites allow the acquisition of high resolution satellite imagery of Korean territory and enables the transition from the conventional cultivation area survey method to new image based methods adopted in advanced nations. In this study, we suggested reasonable sizes of the primary sampling unit and the secondary sampling unit for the satellite imagery based sampling design in 8 provinces preselected for this research. The PSU size was determined mainly in consideration of intracorrelation that shows the degree of homogeneity within each cluster and the efficiency of the image process. For the SSU size, we considered the relative standard error and the differences between the land cover maps produced by the Ministry of Environment and the satellite imagery processed by the National Statistical Office.

Researchers using various of statistical data want to obtain microdata for a detailed analysis. Institutes need to provide microdata after masking processes for sensitive data. Many researchers have used the proportion of unique identity for the measurement of disclosure risk. We proposed a new measurement of disclosure risk that considers the case that all identities are the same or similar. As an application example, we compare the newly proposed measurement and the existing measurement using 10667 data in 'Korea Household Income and Expenditure Survey data for 2010'.

Traditional value at risk(S-VaR) has a difficulity in predicting the future risk of financial asset prices since S-VaR is a backward looking measure based on the historical data of the underlying asset prices. In order to resolve the deficiency of S-VaR, an economic value at risk(E-VaR) using the risk neutral probability distributions is suggested since E-VaR is a forward looking measure based on the option price data. In this study E-VaR is estimated by assuming the generalized gamma distribution(GGD) as risk neutral density function which is implied in the option. The estimated E-VaR with GGD was compared with E-VaR estimates under the Black-Scholes model, two-lognormal mixture distribution, generalized extreme value distribution and S-VaR estimates under the normal distribution and GARCH(1, 1) model, respectively. The option market data of the KOSPI 200 index are used in order to compare the performances of the above VaR estimates. The results of the empirical analysis show that GGD seems to have a tendency to estimate VaR conservatively; however, GGD is superior to other models in the overall sense.

Several statistical models for bivariate poisson data are suggested and used to analyze 2011 K-league data. Our interest is composed of two purposes: The first purpose is to exploit potential attacking and defensive abilities of each team. Particular, a bivariate poisson model with diagonal inflation is incorporated for the estimation of draws. A joint model is applied to estimate an association between poisson distribution and probability of draw. The second one is to investigate causes on scoring time of goals and a regression technique of recurrent event data is applied. Some related future works are suggested.

In many sports games, we would use a coin or the game of rock-paper-scissors to decide which side will begin rst. We consider the game of rock-paper-scissors when two teams are composed of two individuals respectively. There are many methods to choose one team of the two. We consider all 3 cases when all 4 individuals participate simultaneously or one by one. According to the methods of the game rules we nd the means and variances of the number of games respectively.

One proposal is made to construct a nonparametric estimator of slope parameters in a regression model under symmetric error distributions. This estimator is based on the use of the idea of minimizing approximate variance of a proposed estimator using regression quantiles. This nonparametric estimator and some other L-estimators are studied and compared with well known M-estimators through a simulation study.

For credit evaluation models, we extend the study of discriminatory power based on AUC obtained from a ROC curve when the number of defaults is small and distribution functions of the defaults and non-defaults are normal distributions. Since distribution functions do not satisfy normality in real world, the distribution functions of the defaults and non-defaults are assumed as normal mixture distributions based on results that the normal mixture could be better fitted than other distribution estimation methods for non-normal data. By using several AUC statistics, the discriminatory power under such a circumstance is explored and compared with those of normal distributions.

A surplus process with two types of claims is considered, where Type I claims occur more frequently, however, their sizes are smaller stochastically than Type II claims. The ruin probabilities of the surplus caused by each type of claim are obtained by establishing integro-differential equations for the ruin probabilities. The formulas of the ruin probabilities contain an infinite sum and convolutions that make the formulas hard to be applicable in practice; subsequently, we obtain explicit formulas for the ruin probabilities when the sizes of both types of claims are exponentially distributed. Finally, we show through a numerical example, that Type II claims have more impact on the ruin probability of the surplus than Type I claims.

In this study, we propose a new nonparametric test procedure for the multivariate data. In order to accommodate the generalized alternatives for the multivariate case, we construct test statistics via-values with some useful combining functions. Then we illustrate our procedure with an example and compare efficiency among the combining functions through a simulation study. Finally we discuss some interesting features related with the new nonparametric test as concluding remarks.

When the input features are generated by factors in a classification problem, it is more meaningful to identify important factors, rather than individual features. The $F_{\infty}$-norm support vector machine(SVM) has been developed to perform automatic factor selection in classification. However, the $F_{\infty}$-norm SVM may suffer from estimation inefficiency and model selection inconsistency because it applies the same amount of shrinkage to each factor without assessing its relative importance. To overcome such a limitation, we propose the adaptive $F_{\infty}$-norm ($AF_{\infty}$-norm) SVM, which penalizes the empirical hinge loss by the sum of the adaptively weighted factor-wise $L_{\infty}$-norm penalty. The $AF_{\infty}$-norm SVM computes the weights by the 2-norm SVM estimator and can be formulated as a linear programming(LP) problem which is similar to the one of the $F_{\infty}$-norm SVM. The simulation studies show that the proposed $AF_{\infty}$-norm SVM improves upon the $F_{\infty}$-norm SVM in terms of classification accuracy and factor selection performance.

Monte Carlo methods have been used in exact inference for contingency tables for a long time; however, they suffer from ergodicity and the ability to achieve a desired proportion of valid tables. In this paper, we apply the stochastic approximation Monte Carlo(SAMC; Liang et al., 2007) algorithm, as an adaptive Markov chain Monte Carlo, to the exact test of mutual independence in a multiway contingency table. The performance of SAMC has been investigated on real datasets compared to with existing Markov chain Monte Carlo methods. The numerical results are in favor of the new method in terms of the quality of estimates.

This article is concerned with choice-based conjoint analysis versus rating-based and ranking-based conjoint analysis. Choice-based conjoint analysis has a definite advantage in that the respondent's task of choosing the most preferred profile from several competing profiles adequately mimics consumer marketplace behavior. It is crucial to design the choice sets appropriate for the choice-based conjoint. Thus, this article suggests a new method to design the choice sets that are well-balanced. It augments the balanced incomplete block design and then obtains the dual design of the result to accommodate various numbers of profiles. In consequence, the choice sets designed by the new method have the desirable characteristics that each profile is presented to the same number of respondents, and pairs of any two distinct profiles occur together in the same number of choice sets. The balancing of the design increases the efficiency of the conjoint analysis. In addition, the pair-comparison scheme can improve the quality of data through the identification of contradictory responses.

Poisson regression and negative binomial regression are usually used to analyze counting data; however, these models are unsuitable for fit zero-inflated data that contain unexpected zero-valued observations. In this paper, we review the zero-inflated regression in which Bernoulli process and the counting process are hierarchically mixed. It is known that zero-inflated regression can efficiently model the over-dispersion problem. Vuong statistic is employed to compare performances of the zero-inflated models with other standard models.

In this study, we identify the variables that affect the number of crime and spatial correlation in the Seoul metropolitan area, in addition, we measure the relative risk on the incidence of crime by a Poisson regression model. We suggest a statistical methodology to make a risk map for crime based on relative risk instead of the total event of crime by region using the Geographic Information System. To demonstrate the use and advantages of this methodology, this study presents an analyses of the total crime count in 25 wards in the Seoul metropolitan area.