• The Korean Journal of Applied Statistics
• /
• v.35 no.1
• /
• pp.131-146
• /
• 2022

Logit models are commonly used to predicting and classifying categorical response variables. Most Bayesian approaches to logit models are implemented based on the Metropolis-Hastings algorithm. However, the algorithm has disadvantages of slow convergence and difficulty in ensuring adequacy for the proposal distribution. Therefore, we use auxiliary mixture sampler proposed by Frühwirth-Schnatter and Frühwirth (2007) to estimate logit models. This method introduces two sequences of auxiliary latent variables to make logit models satisfy normality and linearity. As a result, the method leads that logit model can be easily implemented by Gibbs sampling. We applied the proposed method to diabetes data from the Community Health Survey (2020) of the Korea Disease Control and Prevention Agency and compared performance with Metropolis-Hastings algorithm. In addition, we showed that the logit model using auxiliary mixture sampling has a great classification performance comparable to that of the machine learning models.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.28 no.1B
• /
• pp.55-64
• /
• 2008

The evidence of changes in the climate system is obvious in the world. Nevertheless, at the current techniques for flood frequency analysis, the flood distribution can not reflect climate change or long-term climate cycles. Using a linear regression and a Mann-Kendall test, trends in annual maximum precipitation and flood data for several major gauging sites were evaluated. Moreover, this research considered incorporating flood trends by climate change effects in flood frequency analyses. For five rainfall gauging sites (Seoul, Incheon, Ulleungdo, Jeonju, and Gangneung), upward trends were observed in all gauged annual maximum precipitation records but they were not statistically significant. For three streamflow gauging sites (Andong Dam, Soyanggang Dam, and Daecheong Dam), upward trends were also observed in all gauged annual maximum flood records, but only the flood at Andong Dam was statistically significant. A log-normal trend model was introduced to reflect the observed linear trends in annual maximum flood series and applied to estimate flood frequency and risk for Andong Dam and Soyanggang Dam. As results, when the target year was 2005, 50-year floods of the log-normal trend model were 41% and 21% larger then those of a log-normal model for Andong Dam and Soyanggang Dam, respectively. Moreover, the estimated floods of the log-normal trend model increases as the target year increases.

• The Korean Journal of Applied Statistics
• /
• v.32 no.4
• /
• pp.631-642
• /
• 2019

Many methods have been developed to solve problems found in sample surveys involving a large number of item non-responses that cause inaccuracies in estimation. However, the non-response adjustment method used under the assumption of random non-response generates a bias in cases where the response rate is affected by the variable of interest. Chung and Shin (2017) and Min and Shin (2018) proposed a method to improve the accuracy of estimation by appropriately adjusting a bias generated when the response rate is a function of the variables of interest. In this study, we studied a case where the response rate function is linear and the error of the super population model follows normal distribution. We also examined the effect of the number of stratum population on bias adjustment. The performance of the proposed estimator was examined through simulation studies and confirmed through actual data analysis.

• The Korean Journal of Applied Statistics
• /
• v.34 no.5
• /
• pp.697-710
• /
• 2021

Relative error prediction is preferred over ordinary prediction methods when relative/percentile errors are regarded as important, especially in econometrics, software engineering and government official statistics. The relative error prediction techniques have been developed in linear/nonlinear regression, nonparametric regression using kernel regression smoother, and stationary time series models. However, random effect models have not been used in relative error prediction. The purpose of this article is to extend relative error prediction to some of generalized linear mixed model (GLMM) with panel data, which is the random effect models based on gamma, lognormal, or inverse gaussian distribution. For better understanding, the real auto insurance data is used to predict the claim size, and the best predictor and the best relative error predictor are comparatively illustrated.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.5 no.4
• /
• pp.406-413
• /
• 1993

Among 2-dimensional coastal hydrodynamic finite element models time stepping ADCIRC and STEPM. and harmonic FUNDY and TEA models were compared in order to find out their characteristics and analyze ernr. General feasibility and capability of models were studied by comparing model results with an analytical solution on some reference points and L$_2$norm error in quarter annular domain where analytical solution can be obtained. According to these tests harmonic models FUNDY and TEA were nearly coinciding with analytical solutions and gave better results than time stepping models. STEPM was at least 5 times better than ADCIRC in L$_2$norm error test and it was 7 times worse than harmonic models. It was expected and concluded that these errors might come from phase lag due to cold start condition and nonlinear effect in basic equations of time stepping models.

• The Korean Journal of Financial Management
• /
• v.21 no.2
• /
• pp.211-233
• /
• 2004

This purpose of paper is to propose a European option pricing formula when the rate of return follows the leptokurtic distribution instead of normal. This distribution explains well the volatility smile and furthermore the option prices calculated under the leptokurtic distribution are shown to be closer to the market prices than those of Black-Scholes model. We make an estimation of the implied volatility and kurtosis to verify the fitness of the pricing formula that we propose here.

• Journal of the Korean Data and Information Science Society
• /
• v.27 no.4
• /
• pp.959-967
• /
• 2016

Distributional assumptions on equity returns play a key role in valuation theories for derivative securities. Elberlein and Keller (1995) investigated the distributional form of compound returns and found that some of standard assumptions can not be justified. Instead, Generalized Hyperbolic (GH) distribution fit the empirical returns with high accuracy. Hu and Kercheval (2007) also show that the normal distribution leads to VaR (Value at Risk) estimate that significantly underestimate the realized empirical values, while the GH distributions do not. We consider saddlepoint approximations to estimate the VaR and the ES (Expected Shortfall) which frequently encountered in finance and insurance as measures of risk management. We supposed GH distributions instead of normal ones, as underlying distribution of linear portfolios. Simulation results show the saddlepoint approximations are very accurate than normal ones.

• Environmental and Resource Economics Review
• /
• v.20 no.1
• /
• pp.119-153
• /
• 2011

This study reviews various aspects of model formulating processes of dichotomous choice responses of the contingent valuation method (CVM), which has been increasingly used in the preliminary feasibility test of Korea public investment projects. The theoretical review emphasizes the consistency between WTP estimation process and WTP measurement process. The empirical analysis suggests that two common parametric models for dichotmous choice responses (RUM and RWTP) and two commonly used probability distributions of random components (probit and logit) resulted in all most the same empirical WTP distributions, as long as the WTP functions are specified to be a linear function of the bid amounts. However, the efficiency gain of DB response compared to SB response were supported on the ground that the two CV responses are derived from the same WTP distribution. Moreover for the exponential WTP function which guarantees the non-negative WTP measures, sample mean WTP were quite different from median WTP if the scale parameter of WTP function turned out to be large.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.20 no.2
• /
• pp.200-209
• /
• 2019

When forecasting future values, a model estimated after minimizing training errors can yield test errors higher than the training errors. This result is the over-fitting problem caused by an increase in model complexity when the model is focused only on a given dataset. Some regularization and resampling methods have been introduced to reduce test errors by alleviating this problem but have been designed for use with only a given dataset. In this paper, we propose a new optimization approach to reduce test errors by transforming a test error minimization problem into a training error minimization problem. To carry out this transformation, we needed additional data for the given dataset, termed pseudo data. To make proper use of pseudo data, we used three types of missing data imputation techniques. As an optimization tool, we chose the least squares method and combined it with an extra pseudo data instance. Furthermore, we present the numerical results supporting our proposed approach, which resulted in less test errors than the ordinary least squares method.

• The Korean Journal of Financial Management
• /
• v.26 no.3
• /
• pp.139-169
• /
• 2009

This paper introduces the modified VaR which takes into account the asymmetry and fat-tails of financial asset distribution, and then compares its out-of-sample forecast performance with traditional VaR model such as historical simulation model and Riskmetrics. The empirical tests using stock indices of 6 countries showed that the modified VaR has the best forecast accuracy. At the test of independence, Riskmetrics and GARCH model showed best performances, but the independence was not rejected for the modified VaR. The Monte Carlo simulation using skew t distribution again proved the best forecast performance of the modified VaR. One of many advantages of the modified VaR is that it is appropriate for measuring VaR of the portfolio, because it can reflect not only the linear relationship but also the nonlinear relationship between individual assets of the portfolio through coskewness and cokurtosis. The empirical analysis about decomposing VaR of the portfolio of 6 stock indices confirmed that the component VaR is very useful for the re-allocation of component assets to achieve higher Sharpe ratio and the active risk management.