• Journal of the Korean Statistical Society
• /
• 제28권2호
• /
• pp.151-165
• /
• 1999

This paper considers the trimmed least squares estimator of the autoregression parameter in the unstable AR(1) model: X\ulcorner=ØX\ulcorner+$\varepsilon$\ulcorner, where $\varepsilon$\ulcorner are iid random variables with mean 0 and variance $\sigma$$^2$> 0, and Ø is the real number with │Ø│=1. The trimmed least squares estimator for Ø is defined in analogy of that of Welsh(1987). The limiting distribution of the trimmed least squares estimator is derived under certain regularity conditions.

• Journal of applied mathematics & informatics
• /
• 제4권2호
• /
• pp.501-512
• /
• 1997

In multiple linear regression model we have presupposed assumptions (independence normality variance homogeneity and so on) on error term. When case weights are given because of variance heterogeneity we can estimate efficiently regression parameter using weighted least squares estimator. Unfortunately this estimator is sen-sitive to outliers like ordinary least squares estimator. Thus in this paper we proposed some statistics for detection of outliers in weighted least squares regression.

• Communications for Statistical Applications and Methods
• /
• 제10권3호
• /
• pp.1037-1046
• /
• 2003

We propose an equivariant and robust estimator in multivariate regression model based on the least trimmed squares (LTS) estimator in univariate regression. We call this estimator as multivariate least trimmed squares (MLTS) estimator. The MLTS estimator considers correlations among response variables and it can be shown that the proposed estimator has the appropriate equivariance properties defined in multivariate regression. The MLTS estimator has high breakdown point as does LTS estimator in univariate case. We develop an algorithm for MLTS estimate. Simulation are performed to compare the efficiencies of MLTS estimate with coordinatewise LTS estimate and a numerical example is given to illustrate the effectiveness of MLTS estimate in multivariate regression.

• 전기학회논문지
• /
• 제65권9호
• /
• pp.1541-1550
• /
• 2016

In this paper, we introduce a novel learning methodology of fuzzy clustering-based neural network pattern classifier. Fuzzy clustering-based neural network pattern classifier depicts the patterns of given classes using fuzzy rules and categorizes the patterns on unseen data through fuzzy rules. Least squares estimator(LSE) or weighted least squares estimator(WLSE) is typically used in order to estimate the coefficients of polynomial function, but this study proposes a novel coefficient estimate method which includes advantages of the existing methods. The premise part of fuzzy rule depicts input space as "If" clause of fuzzy rule through fuzzy c-means(FCM) clustering, while the consequent part of fuzzy rule denotes output space through polynomial function such as linear, quadratic and their coefficients are estimated by the proposed local least squares estimator(LLSE)-based learning. In order to evaluate the performance of the proposed pattern classifier, the variety of machine learning data sets are exploited in experiments and through the comparative analysis of performance, it provides that the proposed LLSE-based learning method is preferable when compared with the other learning methods conventionally used in previous literature.

• Journal of the Korean Statistical Society
• /
• 제31권2호
• /
• pp.251-259
• /
• 2002

Lagged regressor models with general stationary errors independent of the regressors are considered. The regressor process is unstable having characteristic roots on the unit circle. If the order of the lag matches the number of roots on the unit circle, the ordinary least squares estimator (OLSE) is asymptotically efficient in that it has the same limiting distribution as the generalized least squares estimator (GLSE) under the same normalization. This result extends the well-known result of Grenander and Rosenblatt (1957) for asymptotic efficiency of the OLSE in deterministic polynomial and/or trigonometric regressor models to a class of models with stochastic regressors.

• Journal of the Korean Statistical Society
• /
• 제28권1호
• /
• pp.45-55
• /
• 1999

Knowing the number of unit roots is important in the analysis of k-dimensional multivariate autoregressive process. In this paper we suggest simple multiple unit roots test statistics based on least squares estimator for the multivariate AR(1) process in which some eigenvalues are one and the rest are less than one in magnitude. The empirical distributions are tabulated for suggested test statistics. We have small Monte-Calro studies to compare the powers of the test statistics suggested by Johansen(1988) and in this paper.

• 응용통계연구
• /
• 제6권1호
• /
• pp.79-93
• /
• 1993

회귀모형에 있어서의 Ruppert와 Carroll의 절사 회귀 추정법을 확장하여 회귀 분위수에 의 한 두 개의 두분으로 관측치를 분할하여 각 부분마다 가중치를 달리 부여하는 방법으로 적 합된 L-추정법을 제안하였다. 이 제안된 L-추정법은 특히 비대칭인 오차분포하에서 좋은 효율을 가지고 있었다.

• Journal of the Korean Statistical Society
• /
• 제8권1호
• /
• pp.23-36
• /
• 1979

In the multiple linear regression model a class of weighted least absolute error estimaters, which minimize the sum of weighted absolute residuals, is proposed. It is shown that the weighted least absolute error estimators with Wilcoxon scores are equivalent to the Koul's Wilcoxon type estimator. Therefore, the asymptotic efficiency of the proposed estimator with Wilcoxon scores relative to the least squares estimator is the same as the Pitman efficiency of the Wilcoxon test relative to the Student's t-test. To find the estimates the iterative weighted least squares method suggested by Schlossmacher is applicable.

• Communications for Statistical Applications and Methods
• /
• 제8권3호
• /
• pp.633-641
• /
• 2001

The least squares method is usually applied when estimating the parameters in the regression models. However the least square estimator is not very efficient when the distribution of the error is skewed. In this paper, we propose the asymmetric least square estimator for a particular nonlinear time series regression model, and give the simple and practical sufficient conditions for the strong consistency of the estimators.

• Journal of the Korean Data and Information Science Society
• /
• 제17권4호
• /
• pp.1169-1180
• /
• 2006

Partial least squares regression (PLS) is a biased, non-least squares regression method and is an alternative to the ordinary least squares regression (OLS) when predictors are highly collinear or predictors outnumber observations. One way to understand the properties of biased regression methods is to know how the estimators shrink the OLS estimator. In this paper, we introduce an expression for the shrinkage factor of PLS and develop a new shrinkage expression, and then prove the equivalence of the two representations. We use two near-infrared (NIR) data sets to show general behavior of the shrinkage and in particular for what eigendirections PLS expands the OLS coefficients.