• The Korean Journal of Applied Statistics
• /
• v.13 no.2
• /
• pp.541-562
• /
• 2000

The generalized linear mixed model framework is for handling count-type categorical data as well as for clustered or overdispersed non-Gaussian data, or for non-linear model data. In this study, we review its general formulation and estimation methods, based on quasi-likelihood and Monte-Carlo techniques. The current research areas and topics for further development are also mentioned.

• The Korean Journal of Applied Statistics
• /
• v.18 no.2
• /
• pp.381-393
• /
• 2005

As a promising technique for dimension reduction in regression analysis, Sliced Inverse Regression (SIR) and an associated chi-square test for dimensionality were introduced by Li (1991). However, Li's test needs assumption of Normality for predictors and found to be heavily dependent on the number of slices. We will provide a unified asymptotic test for determining the dimensionality of the SIR model which is based on the probabilistic principal component analysis and free of normality assumption on predictors. Illustrative results with simulated and real examples will also be provided.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2018.05a
• /
• pp.303-303
• /
• 2018

안정적인 수자원 운용을 위해서는 정확한 유량예측 기술이 필요하다. 본 연구에서는 유량예측 정확도의 개선을 위해 베이지안 추론(Bayesian inference) 기법과 앙상블 유량 예측(Ensemble Streamflow Prediction, ESP) 기법의 결합을 통한 새로운 유량예측 기법(Bayesian ESP)을 제안하였다. ESP를 통한 유량 예보 앙상블은 베이지안 추론의 사전정보로 활용되며, 관측 유량과 ESP 전망 결과의 선형관계를 통해 우도함수가 추정된다. 우도함수는 관측 유량이 존재하는 과거 기간에 대한 ESP를 수행한 후 예보 시점의 관측 유량(concurrent observed flow)과 선행 관측 유량(lagged observed flow)과의 다중선형회귀 모형을 통해 추정된다. 사전정보와 우도함수는 정규분포로 가정되며, 따라서 최종 유량예측인 사후정보 역시 정규분포함수로 산정되게 된다. Bayesian ESP은 ESP에서 발생하는 강우-유출모형 오차의 개선을 통해 수문예측의 정확도를 개선하게 되며 정규분포함수로 최종 결과가 산정되므로 확률예보 형태의 수문 전망도 가능하다. 본 기법을 전국 35개 댐 유역에 시범적용을 한 결과, 모든 유역에서 기존 ESP 기법 대비 수문예측 정확도의 개선을 가져왔으며, 우도함수 추정에 있어 선행 유량의 포함 여부가 수문 예측 정확도의 추가적인 개선을 가져왔다. 본 기법은 주간 예보부터 계절 예보까지 탄력적으로 구축이 가능하며 적용 결과 리드 타임이 길어질수록 예측 능력이 감소되었지만 전체 구간에 있어서 Bayesian ESP 기법이 가장 우수한 예측 정확도를 보여주었다.

• The Korean Journal of Applied Statistics
• /
• v.33 no.1
• /
• pp.11-23
• /
• 2020

Sampling design should provide statistics to meet a given accuracy while saving cost and time. However, a large number of non-responses are occurring due to the deterioration of survey circumstances, which significantly reduces the accuracy of the survey results. Non-responses occur for a variety of reasons. Chung and Shin (2017, 2019) and Min and Shin (2018) found that the accuracy of estimation is improved by removing the bias caused by non-response when the response rate is an exponential or linear function of variable of interests. For that case they assumed that the error of the super population model follows normal distribution. In this study, we proposed a non-response bias adjusted estimator in the case where the error of a super population model follows the gamma distribution or the log-normal distribution in a business survey. We confirmed the superiority of the proposed estimator through simulation studies.

• The Korean Journal of Applied Statistics
• /
• v.26 no.6
• /
• pp.953-958
• /
• 2013

This study suggests a measurement error model based on skewed normal distribution instead of normal distribution to identify slope parameter properties in a simple liner regression model. We prove that the slope parameter in a simple linear regression model is underestimated.

• The Journal of Information Technology
• /
• v.3 no.1
• /
• pp.37-44
• /
• 2000

The purchasing estimate price of diamond is affected by the factors of carat, color, clarity, certificate, cut and price with the unit of $/carat. The object of this study is to obtain the linear regression model for such purchasing estimate price and to test statistically. The optimum model is the simple regression model of $^y{\;}:{\;}10^2{\;}/{\;}(-1.5575{\;}+{\;}0.3099{\;}logx){\;}+{\;}{\varepsilon}$ statistically satisfied by the lack of fit test and has the characteristics of normality, constant variance and symmetry.

• The Korean Journal of Applied Statistics
• /
• v.17 no.2
• /
• pp.219-228
• /
• 2004

In this paper we construct prediction intervals for nonlinear time series models using the bootstrap. We compare these prediction intervals to traditional asymptotic prediction intervals using quasi-score estimation function and M-quasi-score estimating function comprising bounded functions. Simulation results show that the bootstrap method leads to improved accuracy. The accuracy of the bootstrap is empirically demonstrated with the consumer price index.

• Communications for Statistical Applications and Methods
• /
• v.15 no.1
• /
• pp.27-41
• /
• 2008

A graphical method of checking the adequacy of a generalized linear model is proposed. The graph helps to assess the assumption that the link function of mean can be expressed as a linear combination of explanatory variables in the generalized linear model. For the graph the boosting technique is applied to estimate nonparametrically the relationship between the link function of the mean and the explanatory variables, though any other nonparametric regression methods can be applied. Through simulation studies with normal and binary data, the effectiveness of the graph is demonstrated. And we list some limitations and technical details of the graph.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.11 no.2
• /
• pp.169-174
• /
• 2018

In general, the methods of the analysis of variance(ANOVA) for the continuous data and the chi-square test for the discrete data are used for statistical analysis of the effect and the association. In multidimensional data, analysis of hierarchical structure is required and statistical linear model is adopted. The structure of the linear model requires the normality of the data. A multidimensional categorical data analysis methods are used for causal relations, interactions, and correlation analysis. In this paper, Bayesian network model using probability distribution is proposed to reduce analysis procedure and analyze interactions and causal relationships in categorical data analysis.

• Journal of Korea Water Resources Association
• /
• v.39 no.12 s.173
• /
• pp.997-1012
• /
• 2006

We have presented a nonparametric stochastic approach for the SOI(Southern Oscillation Index) series that used nonlinear methodology called Nonlinear AutoRegressive(NAR) based on conditional kernel density function and CAFPE(Corrected Asymptotic Final Prediction Error) lag selection. The fitted linear AR model represents heteroscedasticity, and besides, a BDS(Brock - Dechert - Sheinkman) statistics is rejected. Hence, we applied NAR model to the SOI series. We can identify the lags 1, 2 and 4 are appropriate one, and estimated conditional mean function. There is no autocorrelation of residuals in the Portmanteau Test. However, the null hypothesis of normality and no heteroscedasticity is rejected in the Jarque-Bera Test and ARCH-LM Test, respectively. Moreover, the lag selection for conditional standard deviation function with CAFPE provides lags 3, 8 and 9. As the results of conditional standard deviation analysis, all I.I.D assumptions of the residuals are accepted. Particularly, the BDS statistics is accepted at the 95% and 99% significance level. Finally, we split the SOI set into a sample for estimating themodel and a sample for out-of-sample prediction, that is, we conduct the one-step ahead forecasts for the last 97 values (15%). The NAR model shows a MSEP of 0.5464 that is 7% lower than those of the linear model. Hence, the relevance of the NAR model may be proved in these results, and the nonparametric NAR model is encouraging rather than a linear one to reflect the nonlinearity of SOI series.