非線型回歸模型의 攪亂項表記와 模型選擇에 대한 實證的 硏究 : 生産函數에 대한 適用 = (An) Empirical Study to a Disturbance Term Specification and Model Choice of Nonlinear Regression Models

다국어 초록 (Multilingual Abstract)

In estimating nonlinear regression models, the "systematic part" of the model is specified with reference to the stochastic part(disturbance term). The disturbance term variable is simply introduced at the estimation in one manner. Really, however, we don't know whether the introduced disturbance term specification is good or not. This paper assumes that the disturbance term is introduced with the additive form and with the multiplicative form. And two forms are compared with the likelihood. The criterion of the choice between the additive form and the multiplicative form is that a form with greater likelihood will be selected. The used method is Box-Cox transformation. The nonlinear regression model used in empirical study is the Cobb-Douglas production function and the transcendental production function and the used data is Korea Minor and Manufacturing data.
And this paper considered the model choice of nonlinear regression models. For the model choice, it uses the test method of nested model and the test statistics is the log-likelihood ratio. In empirical study, Cobb-Douglas production function and transcendental production function with the additive disturbance terms are choosed. The result of choice is the transcendental production function.

목차 (Table of Contents)

• 목차 = ⅰ
• 第1章 序論 = 1
• 第2章 非線型回歸樸型의 攪亂項表記와 模型選擇에 대한 理論的 考察 = 4
• 第1節 攪亂項表記의 檢定方法에 대한 理論約 考察 = 4
• 第2節 模型選擇(model choice)의 檢定方法에 대한 理論的 考察 = 12
• 第3章 實證的 分析 = 15
• 第1節 이용한 具體的 形態의 生産函數 = 15
• 第2節 이용한 統計資料 = 17
• 第3節 實證的 分析의 結果 = 18
• 第4章 結論 = 23
• 附錄 = 25
• SUMMARY = 26
• 參考文獻 = 28