원문정보
초록
영어
Firstly, simulation studies on the type I error probability and power of non-central skew-F testing statistic are given by using Monte Carlo method in this paper. Then, the characteristics of the skew-normal distribution are showed on the data of energy intensity in China. Also, skew-normal multivariate regression models are established to study the main influence factors of energy intensity for the whole country, eastern China and mid- western China respectively. On this basis, a GEE model is constructed to verify the robustness of the above skew-normal multivariate regression results. It shows that the technology progress, industry structure and energy consumption structure have significant influences on the energy intensity for the whole country, eastern and mid- western region, with the R&D input and electricity consumption proportion influencing negatively while the secondary industry proportion influencing positively. Moreover, the impacts of technology progress and energy consumption structure turn out to be quite different among regions. Finally, some countermeasures and suggestions are recommended in this paper.
목차
1. Introduction
2. Skew-normal Multivariate Regression Model
2.1. Skew-normal Distribution, Non-central Skew Chi-square Distribution and Non-central Skew-F Distribution
2.2. Skew-normal Multivariate Regression Model
2.3. The Simulation Study of Non-central Skew-F Test Statistic
3. The Empirical Analysis
3.1. Variable Selection and Data Description
3.2. The Verification of Skew-normal Distribution
3.3. The Estimation of Skew-normal Multivariate Regression Model
3.4. Robustness Test: A Comparison between Skew-normal Multivariate Regression Model and GEE
4. Conclusions
Acknowledgements
References