• The Korean Journal of Applied Statistics
• /
• v.32 no.2
• /
• pp.199-211
• /
• 2019

This paper is concerned with issues in the finite mixture of regression modeling as well as the simultaneous selection of the number of mixing components and relevant predictors. We propose a penalized likelihood method for both mixture components and regression coefficients that enable the simultaneous identification of significant variables and the determination of important mixture components in mixture of regression models. To avoid over-fitting and bias problems, we applied smoothly clipped absolute deviation (SCAD) penalties on the logarithm of component probabilities suggested by Huang et al. (Statistical Sinica, 27, 147-169, 2013) as well as several well-known penalty functions for coefficients in regression models. Simulation studies reveal that our method is satisfactory with well-known penalties such as SCAD, MCP, and adaptive lasso.

• Journal of the Korean Data and Information Science Society
• /
• v.20 no.6
• /
• pp.1093-1101
• /
• 2009

Quantile regression has become a more widely used technique to describe the distribution of a response variable given a set of explanatory variables. This paper proposes a novel modelfor quantile regression using doubly penalized kernel machine with support vector machine iteratively reweighted least squares (SVM-IRWLS). To make inference about the shape of a population distribution, the widely popularregression, would be inadequate, if the distribution is not approximately Gaussian. We present a likelihood-based approach to the estimation of the regression quantiles that uses the asymmetric Laplace density.

• The Korean Journal of Applied Statistics
• /
• v.29 no.1
• /
• pp.267-277
• /
• 2016

This paper considers we consider the estimation of copula parameters based on residuals in stochastic regression models. We prove that a semiparametric estimator using residual empirical distributions is consistent under some conditions and apply the results to the copula-ARMA model. We provide simulation results for illustration.

• Journal of Korea Technology Innovation Society
• /
• v.13 no.2
• /
• pp.231-251
• /
• 2010

The numbers of SCI paper or patent in science and technology are expected to be related with the number of researcher and knowledge stock (R&D stock, paper stock, patent stock). The results of the regression model showed that severe multicollinearity existed and errors were made in the estimation and testing of regression coefficients. To solve the problem of multicollinearity and estimate the effect of the independent variable properly, principal component regression model were applied for three cases with S&T knowledge production. The estimated principal component regression function was transformed into original independent variables to interpret properly its effect. The analysis indicated that the principal component regression model was useful to estimate the effect of the highly correlate production factors and showed that the number of researcher, R&D stock, paper or patent stock had all positive effect on the production of paper or patent.

• The Korean Journal of Applied Statistics
• /
• v.31 no.2
• /
• pp.303-313
• /
• 2018

This study considers time series models to forecast drug expenditures in national health insurance. We adopt autoregressive error model (ARE) and transfer function model (TFM) with segmented level and trends (before and after 2012) in order to reflect drug price reduction in 2012. The ARE has only a segmented deterministic term to increase the forecasting performance, while the TFM explains a causality mechanism of drug expenditure with closely related exogenous variables. The mechanism is developed by cross-correlations of drug expenditures and exogenous variables. In both models, the level change appears significant and the number of drug users and ratio of elderly patients variables are significant in the TFM. The ARE tends to produce relatively low forecasts that have been influenced by a drug price reduction; however, the TFM does relatively high forecasts that have appropriately reflected the effects of exogenous variables. The ARIMA model without the exogenous variables produce the highest forecasts.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.10a
• /
• pp.858-860
• /
• 2019

본 논문에서 우리는 범함수 회귀모형을 이용한 양파의 성장단계별 무게를 예측할 수 있는 새로운 통계적 추정방법을 제안한다. 여기서 우리는 풍속, 평균온도, 강우량, 일조량 그리고 습도 등 나타내는 환경요인들을 설명변수들로 사용하고, 양파의 성장단계별 무게를 반응변수로 사용하여 범함수 회귀모형을 적용하였다. 먼저 그래프분석과 상관분석을 통하여 우리는 일일 평균온도는 양파의 무게 증진에 가장 큰 양의상관이 있고, 풍속이나 습도 그리고 일조량들은 양파의 성장에 약간의 영향력이 있으며 강우량은 양파의 성장에 전혀 도움이 안됨을 알 수 있었다. 두 번째로 범함수 회귀 분석을 통하여 얻어진 각 환경요인들에 대한 회귀계수들의 그림을 통하여 우리는 양파의 성장 기간 동안에 이들의 무게를 향상시키기 위해서는 어떻게 환경요인들을 관리해야 되는 가를 알 수 있는 재배방법을 유도하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2017.11a
• /
• pp.770-773
• /
• 2017

Linear Regression 문제를 토대로 변형하여 선형회귀분석, 2차함수모형 회귀분석, '단조 증가(감소)' 3차 함수 모형 회귀분석과 그에 따라 변형되는 gradient descent 알고리즘을 기술한다.

• Communications for Statistical Applications and Methods
• /
• v.15 no.4
• /
• pp.633-642
• /
• 2008

Variable selection algorithm for partial least square regression using penalty function is proposed. We use the fact that usual partial least square regression problem can be expressed as a maximization problem with appropriate constraints and we will add penalty function to this maximization problem. Then simulated annealing algorithm can be used in searching for optimal solutions of above maximization problem with penalty functions added. The HARD penalty function would be suggested as the best in several aspects. Illustrations with real and simulated examples are provided.

• The Korean Journal of Applied Statistics
• /
• v.12 no.1
• /
• pp.109-124
• /
• 1999

일반적으로 오차항이 자기상관되어 있는 선형회귀 모형에서는 회귀계수에 대한 보통최소제곱추정량이 효율적이지 못 하다고 알려져 있다. 그러나 이러한 일반화선형회귀모형에서 독립변수의 형태에 따라서는 OLSE의 사용 가능성을 제시하는 모형이 있다. 본 연구에서는 오차항이 일차 이동평균 과정을 따르는 선형회귀모형에서 여러 추정량들 (GLSE, APX, MAPX)에 대한 OLSE의 상대효율함수를 유도하고 비교 분석하고자 한다. 특히 소표본에서 정확한 상대효율값을 구하여 OLSE의 효율성이 크게 떨어지지 않거나 효율성이 나은 회귀모형들을 제시한다.

• Communications for Statistical Applications and Methods
• /
• v.5 no.3
• /
• pp.777-792
• /
• 1998

병원의 임상연구실험에서 종종 환자들의 치료에 따른 병세의 호전상태를 여러단계로 분류하여 상이한 치료방법에 대한 치료효과간의 차이론 알고자 하는 경우가 있다. 이와 같이 다중상태의 생존자료를 분석하기 위해서 본 논문에서는 semi-Markov 모형에 Cox 회귀모형을 적용하여 회귀계수와 기저생존함수를 추정하고 이를 바탕으로 반응확률함수를 추정하였다. 그리고 본 논문의 결과를 실제 임상실험에서 얻어진 자료에 적용하여 분석하였다.