• Journal of the Korean Statistical Society
• /
• 제30권4호
• /
• pp.551-561
• /
• 2001

This paper deals with the asymptotic properties for statistical inferences of the parameters in nonlinear regression models. As an optimal criterion for robust estimators of the regression parameters, the regression quantile method is proposed. This paper defines the regression quintile estimators in the nonlinear models and provides simple and practical sufficient conditions for the asymptotic normality of the proposed estimators when the parameter space is compact. The efficiency of the proposed estimator is especially well compared with least squares estimator, least absolute deviation estimator under asymmetric error distribution.

• 자원환경지질
• /
• 제28권4호
• /
• pp.433-438
• /
• 1995

The most popular minimization method is based on the least-squares criterion, which uses the $L_2$ norm to quantify the misfit between observed and synthetic data. The solution of the least-squares problem is the maximum likelihood point of a probability density containing data with Gaussian uncertainties. The distribution of errors in the geophysical data is, however, seldom Gaussian. Using the $L_2$ norm, large and sparsely distributed errors adversely affect the solution, and the estimated model parameters may even be completely unphysical. On the other hand, the least-absolute-deviation optimization, which is based on the $L_1$ norm, has much more robust statistical properties in the presence of noise. The solution of the $L_1$ problem is the maximum likelihood point of a probability density containing data with longer-tailed errors than the Gaussian distribution. Thus, the $L_1$ norm gives more reliable estimates when a small number of large errors contaminate the data. The effect of outliers is further reduced by M-fitting method with Cauchy error criterion, which can be performed by iteratively reweighted least-squares method.

• 대한수학회보
• /
• 제36권3호
• /
• pp.451-457
• /
• 1999

This paper provides sufficient conditions which ensure the strong consistency of regression quantiles estimators of nonlinear regression models. The main result is supported by the application of an asymptotic property of the least absolute deviation estimators as a special case of the proposed estimators. some example is given to illustrate the application of the main result.

• Journal of the Korean Statistical Society
• /
• 제27권3호
• /
• pp.349-358
• /
• 1998

Consider the AR(1) model $X_{t}$=$\beta$ $X_{t-1}$+$\varepsilon$$_{t}$ where $\beta$ < 1 is an unknown parameter to be estimated and {$\varepsilon$$_{t}$} denotes the independent and identically distributed error terms with unknown common distribution function F. In this paper, a strong representation for the least absolute deviation (LAD) estimate of $\beta$ in AR(1) models is obtained under some mild conditions on F. on F.F.

• Journal of applied mathematics & informatics
• /
• 제27권1_2호
• /
• pp.315-326
• /
• 2009

The unknown parameters in regression models are usually estimated by using various existing methods. There are several existing methods, such as the least squares method, which is the most common one, the least absolute deviation method, the regression quantile method, and the asymmetric least squares method. For the nonlinear time series regression models, which do not satisfy the general conditions, we will compare them in two ways: 1) a theoretical comparison in the asymptotic sense and 2) an empirical comparison using Monte Carlo simulation for a small sample size.

• 대한수학회논문집
• /
• 제20권1호
• /
• pp.145-159
• /
• 2005

In this paper we consider the problem of testing statistical hypotheses for unknown parameters in nonlinear regression models and propose three asymptotically equivalent tests based on regression quantiles estimators, which are Wald test, Lagrange Multiplier test and Likelihood Ratio test. We also derive the asymptotic distributions of the three test statistics both under the null hypotheses and under a sequence of local alternatives and verify that the asymptotic relative efficiency of the proposed test statistics with classical test based on least squares depends on the error distributions of the regression models. We give some examples to illustrate that the test based on the regression quantiles estimators performs better than the test based on the least squares estimators of the least absolute deviation estimators when the disturbance has asymmetric and heavy-tailed distribution.

• 응용통계연구
• /
• 제23권6호
• /
• pp.1235-1243
• /
• 2010

다차원 분할표를 구성하는 범주형 변수들의 연관관계를 식별하기 위하여 널리 이용되는 로그선형모형을 위한 절사 LAD(least absolute deviations) 추정방법을 제안하였다. 제안된 방법은 가중 LAD 추정을 반복하여 계산이 수행되므로 분할표 분석을 위해 적용할 수 있는 여러 연관성 모형(association models)에 직접 적용할 수 있다. 또한 붓스트랩을 이용한 최적절사율을 결정하는 방법이 갖는 공분산행렬을 과소추정하는 문제를 해결하기위한 절사율 결정 방법을 제안하였다. 모의실험을 통해 제안된 방법이 붓스트랩 방법에 비하여 항상 우수한 절사율을 보인다는 것을 설명하였으며, 제안된 방법들의 실제 자료분석 결과를 제시하였다.

• Communications for Statistical Applications and Methods
• /
• 제25권6호
• /
• pp.591-604
• /
• 2018

The accelerated failure time (AFT) model is a linear model under the log-transformation of survival time that has been introduced as a useful alternative to the proportional hazards (PH) model. In this paper we propose variable-selection procedures of fixed effects in a parametric AFT model using penalized likelihood approaches. We use three popular penalty functions, least absolute shrinkage and selection operator (LASSO), adaptive LASSO and smoothly clipped absolute deviation (SCAD). With these procedures we can select important variables and estimate the fixed effects at the same time. The performance of the proposed method is evaluated using simulation studies, including the investigation of impact of misspecifying the assumed distribution. The proposed method is illustrated with a primary biliary cirrhosis (PBC) data set.

• 응용통계연구
• /
• 제35권5호
• /
• pp.631-644
• /
• 2022

본 논문은 고차원 시계열 자료에 이상점이 존재하는 경우 희박벡터자기회귀모형(sparse VAR; sVAR)의 모수를 강건하게 추정하는 방법에 대해서 연구하였다. 먼저 Xu 등 (2008)이 독립인 자료에서 밝혔듯이 adaptive lasso 방법이 sVAR 모형에서도 어느 정도의 강건함을 가짐을 모의 실험을 통해 알 수 있었다. 하지만, 이상점의 개수가 증가하거나 이상점의 영향력이 커지는 경우 효율성이 현저히 저하되는 현상도 관찰할 수 있었다. 따라서 이를 개선하기 위해서 최소절대편차(least absolute deviation; LAD)와 Huber 함수를 기반으로 벌점화 시키는 adaptive lasso를 이용하여 sVAR 모형을 추정하는 방법을 본 논문에서는 제안하고 그 성능을 검토하였다. 모의 실험을 통해 제안한 로버스트 추정 방법이 이상점이 존재하는 경우에 모수 추정을 더 정확하게 하고 예측 성능도 뛰어남을 확인했다. 또한 해당 방법론들을 전력사용량 데이터에 적용한 결과 이상점으로 의심되는 시점들이 존재하였고, 이를 고려하여 강건하게 추정하는 제안한 방법론이 더 좋은 예측 성능을 보임을 확인할 수 있었다.

• 대한수학회논문집
• /
• 제33권3호
• /
• pp.973-983
• /
• 2018

Fuzzy linear regression model has been widely studied with many successful applications but there have been only a few studies on the fuzzy regression model with monotonic response function as a generalization of the linear response function. In this paper, we propose the fuzzy regression model with the monotonic response function and the algorithm to construct the proposed model by using ${\alpha}-level$ set of fuzzy number and the resolution identity theorem. To estimate parameters of the proposed model, the least squares (LS) method and the least absolute deviation (LAD) method have been used in this paper. In addition, to evaluate the performance of the proposed model, two performance measures of goodness of fit are introduced. The numerical examples indicate that the fuzzy regression model with the monotonic response function is preferable to the fuzzy linear regression model when the fuzzy data represent the non-linear pattern.