项目名称: 半参数混合密度模型的理论及应用
项目编号: No.11301324
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 数理科学和化学
项目作者: 黄勉
作者单位: 上海财经大学
项目金额: 22万元
中文摘要: 非参数和半参数模型及其应用是现代统计学研究的前沿课题,在经济、金融、电子商务等领域得到广泛应用。本项目将尝试研究结合非参数和半参数技术的混合密度模型,并对一些金融市场数据进行实证研究。本课题拟在半参数混合密度模型及相关方法上开展一些原创性和拓展性兼备的研究工作,在现有的研究基础上,进一步研究动态半参数混合密度模型和半参数混合回归模型,以及他们在分析公司收益率和其他变量如公司规模、财务状况、负债等的应用。主要的理论研究内容包括半参数混合密度模型的可识别性问题、估计方法和算法、大样本性质。同时,在本课题的研究中,将通过大量的计算机模拟和数据分析检验所提方法的有效性,并将新方法应用于我国金融市场数据的实证分析。
中文关键词: 混合模型;非参数方法;半参数模型;异质性;可识别性
英文摘要: The non-parametric and semi-parametric model and their applications are on the frontier of modern statistical research. They have been widely used in the fields of economy, finance, and electronic commerce. In this project, we will investigate the non-parametric and semi-parametric methods in mixture density model, and their application in empirical research of financial market data. We will focus on the novel semi-parametric mixture density models, their related models and methods, and some extensions of the work. Based on the existing research, we further investigate the dynamic semi-parametric mixture density model and the semi-parametric mixture of regression models, as well as their application in the analysis of return on capital (ROC) and other related variables such as company size, financial condition, liabilities, etc.. The theoretical research topics include the identification problem of the semi-parametric mixture density model, estimation methods and algorithms, and the asymptotic properties. Extensive numerical simulation and data analysis will be conducted to demonstrate the performance and the effectiveness of the proposed method. The newly proposed methodologies will be illustrated using an empirical analysis of China's financial market data.
英文关键词: mixture model;nonparametric method;semiparametric model;heterogeneity;identifiability