非線型計量模型과 켬퓨터 Algorithm

• 영문명: Non-linear Econometric Model and Computer Algorithms
• 발행기관: 건국대학교 경제경영연구소
• 저자명: Sung Whan Ju(朱星煥)
• 간행물 정보: 『상경연구』제10권, 85~98쪽, 전체 14쪽
• 주제분류: 경제경영 > 경제학
• 파일형태: PDF
• 발행일자: 1985.08.10

영문 초록

In recent years, the advent of advanced computer technology has made it possible for the econometrician to estimate an increasing number of non-linear regression models. Non-linearity arises in many diverse ways in econometric applications. The estimations for nonlinear econometric models are, in some cases, derived by transformation of the model into linear models. Unfortinately, for many applied problems, nonlinear specifications can not be avoided. Perhaps the general non-linear models are used in the estimation of production and coat functions. In order to estimate the parameters in the nonlinear model, the objective function has to be specified. From the objective function we can derive either a sum of squared errors or likelihood function on which we apply a numerical optimization method. To use the nunerical optimigation method, we have to select a robust numerical algorithm which can be applicable to a particular econometric model. Also to find a grobal optimum of the objective function, appropriate initial point should be choosed. For nonlinear models, the properties of estimators and test statistics can only be derived approximately or asymptotically. Test for a general hypothesis can he performed by using the asymptotic x²distribution of the Wald, the Rao or the likelihood ratio test statistic under the condition of normally distributed errors. Although the discussion of techniques and methods for estimation in nonlinear models has substantially improved during recent years, many problems are still unsolued. It is difficult to choose the optimal algorithm out of the set of available techniques for a particalar model. Asymptotic properties in the non-linear model are only valid for large-sample sizes. In practise, we typically work with small or moderate samples. Therefore the small-sample and asymptotic distribution of the estimators or test statistics may differ substantially in practice.

목차

Ⅰ. 序論

Ⅱ. 母數의 推定方法

Ⅲ. 推定値의 計算方法

Ⅳ. 統計的 檢定

Ⅵ. 結論

Selected References

SUMMARY