• Communications for Statistical Applications and Methods
• /
• v.11 no.2
• /
• pp.381-397
• /
• 2004

The marginal likelihood has become an important tool for model selection in Bayesian analysis because it can be used to rank the models. We discuss the marginal likelihood for Poisson regression models that are potentially useful in small area estimation. Computation in these models is intensive and it requires an implementation of Markov chain Monte Carlo (MCMC) methods. Using importance sampling and multivariate density estimation, we demonstrate a computation of the marginal likelihood through an output analysis from an MCMC sampler.

• Communications for Statistical Applications and Methods
• /
• v.23 no.4
• /
• pp.297-319
• /
• 2016

Football match predictions are of great interest to fans and sports press. In the last few years it has been the focus of several studies. In this paper, we propose the Poisson regression model in order to football match outcomes. We applied the proposed methodology to two national competitions: the 2012-2013 English Premier League and the 2015 Brazilian Football League. The number of goals scored by each team in a match is assumed to follow Poisson distribution, whose average reflects the strength of the attack, defense and the home team advantage. Inferences about all unknown quantities involved are made using a Bayesian approach. We calculate the probabilities of win, draw and loss for each match using a simulation procedure. Besides, also using simulation, the probability of a team qualifying for continental tournaments, being crowned champion or relegated to the second division is obtained.

• Communications for Statistical Applications and Methods
• /
• v.5 no.1
• /
• pp.1-18
• /
• 1998

Since 1960s, in which the computer graphics system first appeared, various graphical techniques have been introduced for regression diagnostics and they have been remarkably developed. In particular, animation, one of the dynamic graphical methods which Cook and Weisberg (1989) proposed helps to show the effect of adding variables or observations to a model, or removing them from a model on the regression results. We present the added variable plots (AVP) with animation, which can be used as an optical tool of understanding the affect of some variables or observations on other variables, and the detrended added-variable plots (DAVP) with animation, through which it is possible to find out whether specific variables or observations have an effect on the nonlinearity of other variables or not.

• The Korean Journal of Applied Statistics
• /
• v.24 no.4
• /
• pp.587-595
• /
• 2011

Credit scoring is an objective and automatic system to assess the credit risk of each customer. The logistic regression model is one of the popular methods of credit scoring to predict the default probability; however, it may not detect possible nonlinear features of predictors despite the advantages of interpretability and low computation cost. In this paper, we propose to use a generalized partially linear model as an alternative to logistic regression. We also introduce modern ensemble technologies such as bagging, boosting and random forests. We compare these methods via a simulation study and illustrate them through a German credit dataset.

• Journal of Korean Society for Quality Management
• /
• v.28 no.2
• /
• pp.57-69
• /
• 2000

This paper discusses the local influence approach to the linear regression models with AR(2) errors. Diagnostics for the linear regression models with AR(2) errors are proposed and developed when simultaneous perturbations of the response vector are allowed- That is, the direction of maximum curvature of local influence analysis is obtained by studying the curvature of a surface associated with the overall discrepancy measure.

• Journal of the Korean Statistical Society
• /
• v.27 no.3
• /
• pp.319-329
• /
• 1998

In the linear regression model with an autocorrelated disturbances, the Cochrane-Orcutt estimator (COE) is a well known alternative to the Generalized Least Squares estimator (GLSE). The efficiency of COE has been studied empirically in a Monte Carlo study when the unknown parameters are estimated by maximum likelihood method. In this paper, it is theoretically proved that the COE is shown to be inferior to the GLSE. The comparisons are based on the difference of corresponding information matrices or the ratio of their determinants.

• Journal of the Korean Statistical Society
• /
• v.33 no.4
• /
• pp.367-379
• /
• 2004

This article deals with the nonparametric analysis of longitudinal data when there exist possible correlations among repeated measurements for a given subject. We consider a quasi-likelihood regression model where a transformation of the regression function through a link function is linear in time-varying coefficients. We investigate the local polynomial approach to estimate the time-varying coefficients, and derive the asymptotic distribution of the estimators in this quasi-likelihood context. A real data set is analyzed as an illustrative example.

• Communications for Statistical Applications and Methods
• /
• v.25 no.5
• /
• pp.569-575
• /
• 2018

Pareto distribution is important to analyze data in actuarial sciences, reliability, finance, and climatology. In general, unknown parameters of the Pareto distribution are estimated based on the maximum likelihood method that may yield inadequate inference results for small sample sizes and high percent censored data. In this paper, a new approach based on the regression framework is proposed to estimate unknown parameters of the Pareto distribution under the progressive Type-II censoring scheme. The proposed method provides a new regression type estimator that employs the spacings of exponential progressive Type-II censored samples. In addition, the provided estimator is a consistent estimator with superior performance compared to maximum likelihood estimators in terms of the mean squared error and bias. The validity of the proposed method is assessed through Monte Carlo simulations and real data analysis.

• Journal of Korean Society for Quality Management
• /
• v.15 no.2
• /
• pp.50-54
• /
• 1987

A distribution-free rank test for parallelism of regression lines against ordered alternatives is considered. The proposed test statistic is based on the Kepner-Robinson's transformation. The null distribution of the proposed statistic is the same as that of the Wilcoxon signed rank statistic. But, the proposed procedure can be applied only to four or fewer regression lines. The results of a small-sample Monte Carlo study show that the proposed test is comparable with the parametric test in heavy tailed distributions.

• Communications for Statistical Applications and Methods
• /
• v.28 no.3
• /
• pp.281-294
• /
• 2021

In this paper, we propose an adaptive regression method based on the single-layer neural network structure. We adopt a symmetric activation function as units of the structure. The activation function has a flexibility of its form with a parametrization and has a localizing property that is useful to improve the quality of estimation. In order to provide a spatially adaptive estimator, we regularize coefficients of the activation functions via ℓ₁-penalization, through which the activation functions to be regarded as unnecessary are removed. In implementation, an efficient coordinate descent algorithm is applied for the proposed estimator. To obtain the stable results of estimation, we present an initialization scheme suited for our structure. Model selection procedure based on the Akaike information criterion is described. The simulation results show that the proposed estimator performs favorably in relation to existing methods and recovers the local structure of the underlying function based on the sample.