제3변인이 매개효과와 조절효과를 동시에 갖는 모형에서 제3변인의 왜도가 매개효과의 표준오차 추정에 미치는 영향

이슬,장승민 한국심리학회 2014 한국심리학회지 일반 Vol.33 No.2

A model with a third variable that has both mediation and moderation effects, i.e., the moderating mediator model, can be used to determine whether the third variable has mediation effect, moderation effect, or both in a causal relationship between two variables. Because this model is analyzed base on a moderation model, nonessential collinearity between the third variable and the interaction term may increase the standard error of estimation for the second path of the mediation effect and the increased standard error generates underestimation of statistical significance of mediation effect over and beyond the moderation effect of the third variable. Although researchers may use mean-centering on the predictor and the third variable to decrease the collinearity, degree of the decrease depends on the bivariate normality of the explanatory variables. The current study investigated how much nonessential collinearity and standard error of the estimation for the second path of the mediation effect increased as the amount of deviation from bivariate normality of the explanatory variables increased by manipulating skewness of the third variable. We found that high skewness of the third variable produced substantial amount of nonessential collinearity even with mean-centered variables, which negatively influenced on the estimation of the mediation effect in the moderating mediator model. We also found that increase in sample size attenuated the negative effects. The results suggest that a large number of samples are required in applications of the moderating mediator model with a highly skewed third variable. 두 변인의 인과관계에서 제3변인이 매개효과를 갖는지, 조절효과를 갖는지, 혹은 두 효과를 모두 갖는지 확인하기 위한 목적으로 제3변인이 매개효과와 조절효과를 동시에 갖는 모형(조절매개변인 모형)을 사용할 수 있다. 이 모형은 조절모형에 기반하여 제3변인의 매개효과와 조절효과가 동시에 추정되기 때문에 제3변인과 상호작용항 사이의 비본질적 공선성이 매개효과 2차 경로의 표준오차를 증가시켜 조절효과에 비해 매개효과의 통계적 유의성을 과소평가할 수 있다. 예측변인과 제3변인의 평균중심화를 통해 이 공선성을 감소시킬 수 있지만, 공선성의 감소 정도는 두 설명변인의 분포가 이변량 정규성을 따르는 정도에 따라 달라진다. 본 연구는 조절매개변인 모형에서 제3변인의 왜도를 조작함으로써, 설명변인들의 분포가 이변량 정규성에서 벗어나는 정도에 따라 평균중심화된 제3변인과 상호작용항의 공선성과 매개효과의 2차 경로 회귀계수의 표준오차가 얼마나 증가하는지를 가상실험(simulation) 절차를 통해 확인하였다. 평균중심화에도 불구하고 제3변인의 왜도가 크면 우려할 만한 수준의 비본질적 공선성이 유발됨이 확인되었고 동시에 조절매개변인 모형의 매개효과 추정에 부정적 영향을 미침이 확인되었다. 또한 표본크기를 크게 하는 것이 이 부정적 영향을 줄인다는 것도 확인되었다. 이러한 결과는 조절매개변인 모형을 적용하여 동일한 제3변인의 매개효과와 조절효과를 평가하고자 할 때 제3변인의 왜도가 클 것으로 예상되는 경우 충분한 크기의 표본을 사용하는 것이 필요함을 시사한다.

Optimal Restrictions on Regression Parameters For Linear Mixture Model

Ahn, Jung-Yeon,Park, Sung-Hyun The Korean Statistical Society 1999 Journal of the Korean Statistical Society Vol.28 No.3

Collinearity among independent variables can have severe effects on the precision of response estimation for some region of interest in the experiments with mixture. A method of finding optimal linear restriction on regression parameter in linear model for mixture experiments in the sense of minimizing integrated mean squared error is studied. We use the formulation of optimal restrictions on regression parameters for estimating responses proposed by Park(1981) by transforming mixture components to mathematically independent variables.

Dual Generalized Maximum Entropy Estimation for Panel Data Regression Models

Lee, Jaejun,Cheon, Sooyoung The Korean Statistical Society 2014 Communications for statistical applications and me Vol.21 No.5

Data limited, partial, or incomplete are known as an ill-posed problem. If the data with ill-posed problems are analyzed by traditional statistical methods, the results obviously are not reliable and lead to erroneous interpretations. To overcome these problems, we propose a dual generalized maximum entropy (dual GME) estimator for panel data regression models based on an unconstrained dual Lagrange multiplier method. Monte Carlo simulations for panel data regression models with exogeneity, endogeneity, or/and collinearity show that the dual GME estimator outperforms several other estimators such as using least squares and instruments even in small samples. We believe that our dual GME procedure developed for the panel data regression framework will be useful to analyze ill-posed and endogenous data sets.

점과 직선의 프리미티브를 이용한 호모그래피로부터 외부표정요소 결정

김의명(Kim, Eui Myoung),서홍덕(Seo, Hong Deok) 한국측량학회 2024 한국측량학회지 Vol.42 No.5

If the three-dimensional ground coordinates lie on a two-dimensional plane, projective transformation or homography, can be established between the image and ground coordinates by applying the plane equation to the collinearity condition. Homography refers to mapping by projecting from one plane to another, and is particularly widely used in augmented reality, where the attitude and position of the camera must be determined to map objects in real time as the visual changes. In this study, a method was proposed to determine the exterior orientation parameters of a camera from homography using point and line primitives. As a result of the experiments, the reprojection error using the exterior orientation parameters determined by the four points in the image was found to be 1.104 pixels. When comparing the experimental results using two straight lines and three points or three straight lines and two points in the same image, the experimental results using three straight lines showed relatively high accuracy. In addition, in images with large tilt, the exterior orientation parameters could be determined using four straight lines, but the accuracy was not high when analyzing the reprojection error.

Examining Implications of Collinearity among Formative Measures on Structural Model Estimation

Bongsik Shin,Gimun Kim 한국경영정보학회 2012 한국경영정보학회 학술대회논문집 Vol.2012 No.1

안쪽 분절과 바깥 분절의 변형 및 추가가 네온 색 확산에 미치는 효과

정우현,이일우 한국인지및생물심리학회 2010 한국심리학회지 인지 및 생물 Vol.22 No.1

Five experiments were performed to compare the effects of inner and outer segments on the neon color spreading using the modified Redies-Spillmann figure. Experiment 1 examined the effect of collinearity of inner and outer segments on neon color spreading and experiment 2 compared the effect of continuity. Experiment 3-1 and 3-2, investigated the effects of same space and length of colored segments and outer segments on neon color spreading. Experiment 4 tested how the degree of neon color spreading changed when inner and outer segments were added. The result showed that the impact of outer segments was stronger than inner segments on neon color spreading in collineartity and continuity. However, the effects of identical spaces and lengths were less important than collinearity and continuity on color spreading. In contrast with addition of inner segments, when outer segments were added color spreading were reduced significantly. These results suggest that the centrifugal color spreading is stronger than centripetal spreading and that the subjective contours and the color spreading are caused by separate mechanisms. 변형된 Redies-Spillmann 도형을 자극으로 사용하여 색 분절의 바깥쪽에 있는 분절과 안쪽에 있는 분절이 네온 색 확산에 미치는 효과를 비교해 보았다. 실험 1에서는 안쪽 분절과 바깥쪽 분절의 공선성이 네온 색 확산에 미치는 효과를 살펴보았고 실험 2에서는 연결성의 효과를 비교해 보았다. 실험 3-1과 3-2에서는 바깥쪽 분절과 색 분절의 간격의 동일성과 길이의 동일성이 네온 색 확산에 미치는 효과를 살펴보았다. 실험 4에서는 안쪽 분절과 바깥쪽 분절이 추가되었을 때 네온 색 확산의 정도가 어떻게 달라지는지 알아보았다. 실험 결과 안쪽 분절의 공선성과 연결성의 영향보다는 바깥쪽 분절의 영향이 더 큰 것으로 나타났다. 간격의 동일성과 길이의 동일성은 공선성이나 연결성에 비해 네온 색 확산에 미치는 효과가 미미했다. 안쪽 분절이 추가되었을 때는 네온 색 확산이 거의 영향을 받지 않았으나 바깥쪽 분절이 추가될 경우 네온 색 확산이 감소되는 경향이 나타났다. 이런 결과는 색 확산이 안쪽 윤곽과 바깥쪽 윤곽의 내부로 일어날 때 그 진행 방향이 동일하지 않으며 주관적 윤곽과 네온 색 확산은 별개의 기제에 의해서 만들어짐을 시사한다.

GOES-9 영상의 정밀기하보정을 위한 여러 센서모델 분석

허동석 ( Dong Seok Hur ),김태정 ( Tae Jung Kim ) 대한원격탐사학회 2006 대한원격탐사학회지 Vol.22 No.4

A numerical formula that presents relationship between a point of a satellite image and its ground position is called a sensor model. For precise geolocation of satellite images, we need an error-free sensor model. However, the sensor model based on GOES ephemeris data has some error, in particular after Image Motion Compensation (IMC) mechanism has been turned off. To solve this problem, we investigated three sensor models: collinearity model, direct linear transform (DLT) model and orbit-based model. We applied matching between GOES images and global coastline database and used successful results as control points. With control points we improved the initial image geolocation accuracy using the three models. We compared results from three sensor models. As a result, we showed that the orbit-based model is a suitable sensor model for precise geolocation of GOES-9 images.

The Evaluations of Sensor Models for Push-broom Satellite Sensor

Lee, Suk-Kun,Chang, Hoon Korean Society of Surveying 2004 Korean journal of geomatics Vol.4 No.1

The aim of this research is comparing the existing approximation models (e.g. Affine Transformation and Direct Linear Transformation) with Rational Function Model as a substitute of rigorous sensor model of linear array scanner, especially push-broom sensor. To do so, this research investigates the mathematical model of each approximation method. This is followed by the assessments of accuracy of transformation from object space to image space by using simulated data generated by collinearity equations which incorporate or depict the physical aspects of linear array sensor.

로지스틱회귀모형에서 설명변수의 상대적 중요도 지표에 관한 연구

정윤호,연규필 한국자료분석학회 2017 Journal of the Korean Data Analysis Society Vol.19 No.1

One of the important objectives of regression analysis is to verify the relative importances of predictors in the constructed model. Variable importance is generally measured by the proportion at which a variable contributes on the coefficient of determination. We study on the relative importance measures such as Pratt's product measure, Johnson's relative weight, and Budescu's general dominance index in the perspective of how to derive the definition in the case of logistic regression. Furthermore, we suggest a modified Pratt's index using logistic ridge regression which is commonly utilized when the so called multi-collinearity among variables are considerable. A data analysis regarding the variable importance in logistic regression is conducted and the result shows that the proposed index is practicable in the case of multi-collinearity in that it can derive appropriate importance values compared to the Pratt's measure and produce similar importance ranks to the relative weights. 회귀분석의 주요 목적 중 하나는 설명변수들의 상대적 중요도를 파악하는 것이다. 반응변수가 연속형인 회귀모형에서 설명변수의 중요도는 흔히 결정계수 값을 각 설명변수가 기여하는 부분으로 분해함으로써 중요도를 측정한다. 대표적으로 Pratt의 곱측도, Budescu의 일반우세지수, Johnson의 상대가중치 등이 있다. 본 논문에서는 반응변수가 이항형인 로지스틱회귀모형의 경우에 이러한 지표들이 유사결정계수(pseudo R-squared)를 통해 적절히 정의될 수 있음을 살펴보았다. 특히, 일반우세지수나 상대가중치와는 달리 Pratt의 곱측도는 다중공선성이 클 때 설명변수의 중요도 지표로서 음수값을 도출하는 등의 한계가 있기 때문에 그대로 사용될 수 없음을 고려하여, 로지스틱능형회귀모형(logistic ridge regression)에서의 Pratt 곱측도를 산정하는 방안을 제시하였다. 또한, 사례분석을 통해 로지스틱회귀모형에서의 설명변수 중요도 지표들을 서로 비교하고 제안된 지표의 유용성을 실증하였다. 분석 결과 제안된 지표는 다중공선성이 있는 경우에 기존의 Pratt 곱측도의 단점을 보완하면서 상대가중치와 비슷한 변수 중요도 순위를 나타냄을 확인하였다.

SMIT를 활용한 지진하중을 받는 전단 구조물의 응답모드 특성에 관한 연구

장민우 한국지진공학회 2018 한국지진공학회논문집 Vol.22 No.3

The improvement in computing systems and sensor technologies devotes to conduct data-driven structural health monitoring algorithms for existing civil infrastructures. Despite of the development of techniques, the uncertainty oriented from the measurement results in the discrepancy to the actual structural parameters and let engineers or decision makers hesitate to adopt such techniques. Many studies have shown that the modal identification results can be affected by the uncertainties due to the applied methods and the types of loading. This paper aims to compare the performance of modal identification methods using Structural Modal Identification Toolsuite (SMIT) which has been developed to facilitate multiple identification methods with a user-friendly designed platform. The data fed into SMIT processes three stages for the comprehensive identification including preprocessing, eigenvalue estimation, and post-processing. The seismic and white noise response for shear frame model was obtained from numerical simulation. The identified modal parameters is compared to the actual modal parameters. In order to improve the quality of coherence in identified modal parameters, several hurdles including modal phase collinearity and extended modal amplitude coherence were introduced. Numerical simulation conducted on the 5 dof shear frame model were used to validate the effectiveness of using these parameters.