• Communications for Statistical Applications and Methods
• /
• v.16 no.4
• /
• pp.573-578
• /
• 2009

We prove the consistency of a penalized maximum likelihood estimate proposed by Ahn (2001). The PMLE not only avoids the well-known problem that the ordinary likelihood of the normal mixture model is unbounded for any given sample size, but also removes redundant components.

• The Korean Journal of Applied Statistics
• /
• v.34 no.2
• /
• pp.279-293
• /
• 2021

In this paper, we introduce how to construct sufficient conditions for the oracle property in penalized linear regression model. We give formal definitions of the oracle estimator, penalized estimator, oracle penalized estimator, and the oracle property of the oracle estimator. Based on the definitions, we present a unified way of constructing optimality conditions for the oracle property and sufficient conditions for the optimality conditions that covers most of the existing penalties. In addition, we present an illustrative example and results from the numerical study.

• Journal of the Korean Data and Information Science Society
• /
• v.27 no.2
• /
• pp.499-510
• /
• 2016

It is very important to select relevant variables in regression models for survival analysis. In this paper, we introduce a penalized variable-selection procedure in multi-level frailty models based on the "frailtyHL" R package (Ha et al., 2012). Here, the estimation procedure of models is based on the penalized hierarchical likelihood, and three penalty functions (LASSO, SCAD and HL) are considered. The proposed methods are illustrated with multi-country/multi-center bladder cancer survival data from the EORTC in Belgium. We compare the results of three variable-selection methods and discuss their advantages and disadvantages. In particular, the results of data analysis showed that the SCAD and HL methods select well important variables than in the LASSO method.

• The Korean Journal of Applied Statistics
• /
• v.34 no.3
• /
• pp.411-425
• /
• 2021

Accelerated failure time (AFT) model represents a linear relationship between the log-survival time and covariates. We are interested in the inference of covariate's effect affecting the variation of survival times in the AFT model. Thus, we need to model the variance as well as the mean of survival times. We call the resulting model mean and variance AFT (MV-AFT) model. In this paper, we propose a variable selection procedure of regression parameters of mean and variance in MV-AFT model using penalized likelihood function. For the variable selection, we study four penalty functions, i.e. least absolute shrinkage and selection operator (LASSO), adaptive lasso (ALASSO), smoothly clipped absolute deviation (SCAD) and hierarchical likelihood (HL). With this procedure we can select important covariates and estimate the regression parameters at the same time. The performance of the proposed method is evaluated using simulation studies. The proposed method is illustrated with a clinical example dataset.

• The Korean Journal of Applied Statistics
• /
• v.29 no.6
• /
• pp.1147-1154
• /
• 2016

In this paper, we numerically compare two penalized least square methods, the ${\ell}_0$-penalized method and the fused lasso regression (FLR, ${\ell}_1$ penalization), in finding multiple change points of a signal. We find that the ${\ell}_0$-penalized method performs better than the FLR, which produces many false detections in some cases as the theory tells. In addition, the computation of ${\ell}_0$-penalized method relies on dynamic programming and is as efficient as the FLR.

• The Korean Journal of Applied Statistics
• /
• v.28 no.5
• /
• pp.965-976
• /
• 2015

Determining relevant variables for a regression model is important in regression analysis. Recently, a variable selection methods using a penalized likelihood with various penalty functions (e.g. LASSO and SCAD) have been widely studied in simple statistical models such as linear models and generalized linear models. The advantage of these methods is that they select important variables and estimate regression coefficients, simultaneously; therefore, they delete insignificant variables by estimating their coefficients as zero. We study how to select proper variables based on penalized hierarchical likelihood (HL) in semi-parametric frailty models that allow three penalty functions, LASSO, SCAD and HL. For the variable selection we develop a new function in the "frailtyHL" R package. Our methods are illustrated with breast cancer survival data from the Medical Center at Chonnam National University in Korea. We compare the results from three variable-selection methods and discuss advantages and disadvantages.

• Communications for Statistical Applications and Methods
• /
• v.18 no.6
• /
• pp.837-849
• /
• 2011

A self-organizing network is designed to estimate parameters of normal mixtures. SOMN achieves fast convergence and low possibility of divergence even when sample sizes are small, while PMLE eliminate unnecessary components. The proposed network effectively combines the good properties of SOMN and PMLE. Simulation verifies that the proposed network eliminates unnecessary components in normal mixtures when sample sizes are relatively small.

• The Journal of the Korea Contents Association
• /
• v.8 no.12
• /
• pp.465-472
• /
• 2008

Diverse methods of life guidance adopted by the teacher's subjective standards including corporal punishment have inherent problems. Some schools have introduced and exercised the merit and demerit system, which uses merits and demerits to guide the children for desirable life habits and etiquette. And most of them are done off-line and thus have difficulties with real time reference, statistical process, filling out logs, and management. The merit and demerit management system(MDMS) was developed to support the process, statistics, reference, and authority features as well as card issuing. It also promotes life guidance, personality education, participation of the parents, and further desirable cooperation among the teachers, students, and parents. The system has lots of advantages such as reducing resistance from the students against life guidance rules, enabling the students to check their life guidance status at school, allowing the parents to check how their children are doing at school, increasing efficiency of data management, and taking some burden off the shoulders of the teachers doing statistics. MDMS helps the teachers base corporal punishment on the guidance rules, deter direct punishment on certain body parts, and pursue more systematic, scientific, and human life guidance.

• Journal of Korean Association for Spatial Structures
• /
• v.4 no.4 s.14
• /
• pp.91-98
• /
• 2004

Genetic algorithm is one of the best ways to solve a discrete variable optimization problem. This method is an unconstrained optimization technique, so the constraints are handled in an implicit manner. The most popular way of handling constraints is to transform the original constrained problem into an unconstrained problem, using the concept of penalty function. I present the 3 fitness functions which represent the reject strategy, the penalty strategy, and the combined strategy. I make the design program using the 3 fitness Auctions and it is applied to the design problem of a gable frame and a 2 story 3 span frame.

• Proceedings of the KOR-KST Conference
• /
• 1998.10b
• /
• pp.466-466
• /
• 1998

도로 교통의 혼잡이 나날이 증가되고 있는 현실 상황에서 이를 해결하기 위한 새로운 도로의 무제한적 건설은 정보의 예산절약, 필요한 도로용지 확보의 어려움, 환경오염 문제 등으로 인해 현실적인 한계에 이르렀다. 따라서, 이러한 도로의 혼잡상황에 효과적으로 대처하기 위해서는 승용차를 이용하고자 하는 수요를 대량수송이 가능한 대중교통 이용수요로 전환시켜야 하며, 이를 위해서는 대중교통의 서비스수준 제고 및 운영 관리 체계 등의 개선이 필요하다. 이를 위한 전략적 및 운영적 측면에서의 대중교통계획은 미래 대중교통수요의 정확한 예측을 전제로 하여 수립되며, 이러한 수요의 예측은 필수적으로 현실을 보다 더 정확하게 묘사해 줄 수 있는 통행배정모형을 필요로 한다. 대중교통 통행배정은 규칙적인 배차시간과 정해진 노선을 운행하는 고정서비스 시스템으로 구성되어 있어서 한 링크 상에서도 여러 개의 운행노선을 고려해야 하기 때문에 승용차 통행배정과는 독립적으로 취급되어 왔으며, 이로 인해 그 동안 많은 연구가 선행되어 있지 않은 실정이다. 본 연구는 교통예측 프로그램 중의 하나인 EMME/2에서 사용하고 있는 대중교통수요 통행배정 모형인 최적전략모형(Optimal Strategy Model)의 단점을 보완하기 위한 것이다. 최적전략모형은 수요 배정시, 최적전략에 속하는 경로들에 대해 단순히 운행횟수에 비례하여 수요를 배정함으로 인해서, 예를 들면 운행횟수는 많지만 환승이 많은 경로에 수요를 많이 배정하는 것과 같은 비현실적인 결과가 발생하기도 한다. 본 연구는 이를 개선하기 위해서, 두 가지 대안을 제시했다. 먼저, 노선배정에 우선되는 최적경로 탐색시 환승노드에서의 환승에 대한 벌점을 그 노선의 운행회수에 줌으로써 환승이 많은 경로에 수요의 배정이 적게 되도록 하는 방법과 두 번째로 수요의 배정시 운행횟수가 아닌 목적지까지의 통행시간과 대시시간에 따른 확률적 배분을 통해 기존 모형의 단점을 보완하고자 했다.