缺失数据下半参数变系数EV模型的统计推断

发布时间：2018-03-09 21:29

本文选题：半参数变系数模型　切入点：测量误差　出处：《首都经济贸易大学》2013年硕士论文　论文类型：学位论文

【摘要】：半参数变系数模型中既有参数的分量,又有非参数的分量。所以,半参数变系数模型不仅有参数模型利于解释的特点,还有非参数模型的适应性。参数回归模型、半参数回归模型和变系数模型都是半参数变系数模型的特殊形式。如今,半参数变系数模型已经广泛得应用到了信息科学、金融工程、质量控制、交通工程、能源与坏境、计量经济和生物医学等领域。在现实生活中,我们经常遇到缺失数据、测量误差数据等复杂数据。关于复杂数据的研究,现在已经成为现代统计学研究的热点问题之一。所以,在缺失数据下,研究半参数变系数EV模型的统计推断,具有一定的理论价值和现实意义。本文主要研究的是缺失数据下,半参数变系数EV模型的统计推断问题。通过巧妙得构造了基于借补和纠偏的辅助随机向量,提出了一个基于借补和纠偏的经验似然方法。这与之前方法不同的地方在于,本文构造的基于借补和纠偏的经验对数似然比函数能够服从渐进的标准卡方分布,从而能够推断出了参数分量的置信域。
[Abstract]:There are both parametric and non-parametric components in the semi-parametric variable coefficient model. Therefore, the semi-parametric variable coefficient model not only has the characteristics of the parameter model to be easy to interpret, but also has the adaptability of the non-parametric model and the parametric regression model. Semiparametric regression model and variable coefficient model are special forms of semi-parametric variable coefficient model. Nowadays, semi-parametric variable coefficient model has been widely used in information science, financial engineering, quality control, traffic engineering, energy and bad environment. Econometric and biomedical fields. In real life, we often encounter complex data such as missing data, measuring error data and so on. Research on complex data has become one of the hot issues in modern statistical research. It is of certain theoretical and practical significance to study the statistical inference of the semi-parametric variable coefficient EV model. In this paper, we mainly study the statistical inference problem of the semi-parametric variable coefficient EV model with missing data. An empirical likelihood method based on imputation and correction is proposed, which is different from the previous method in that the empirical logarithmic likelihood ratio function constructed in this paper can satisfy the standard chi-square distribution from progressive. Thus the confidence region of the parameter component can be inferred.
【学位授予单位】：首都经济贸易大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：C81

【参考文献】