선형회귀분석에서 새로운 변수 추가를 위한 3차원 그림

• 영문명: Three Dimensional Plot for Adding New Variables in Linear Regression
• 발행기관: 건국대학교 경제경영연구소
• 저자명: Han Son Seo(徐漢巽) Jeoung Il Kim(金正一)
• 간행물 정보: 『상경연구』제21권, 253~261쪽, 전체 9쪽
• 주제분류: 경제경영 > 경제학
• 파일형태: PDF
• 발행일자: 1996.08.15

영문 초록

This article discusses graphical methods for obtaining the effect of adding explanatory variables, which might be nonlinear, in linear regression. The well-known methods, simple-residual-plot, added-variable-plot, and residual-plus-component-plot can not evaluate the correct effect of adding variables. All three plots lost some part of effect of adding variables. There are occasions that the plots do not indicate linear relation when actual effect of adding variables is linear, and vice versa. As an alternative, Johnson and McCulloch(1987) suggested a new method which uses a locally linear approximation. The locally linear approximation method is extended to the case of two adding variables. Though it has a problem in determining the function of adding variables just through plotting, simulation study indicates that it’ll work well for most cases.

목차

Ⅰ. 서론

Ⅱ. 부분 선형 추정법

Ⅲ. 3차원 추가변수 그림

Ⅳ. 결론

참고문헌

SUMMARY