因子分析模型L的优良性和应用  

The Superiority and Application of The Factor Analysis Model L

作　　者：林海明[1,2] 

机构地区：[1]广东商学院国民经济研究中心 [2]广东商学院华商学院

出　　处：《数量经济技术经济研究》2013年第3期96-113,共18页Journal of Quantitative & Technological Economics

基　　金：教育部人文社会科学研究规划基金项目(2009YJA910002);教育部人文社会科学重点研究基地重大项目(2009JJD910001);广东省普通高校人文社科研究项目(10WYXM020);广东商学院科学研究重点项目(08ZD11001);广东商学院华商学院院级重点项目(HS2011047)的资助

摘　　要：对于传统的因子分析模型,公因子解会不能降维或失去一些变量解释,故因子分析需要找出一个优良模型,而优良模型的确定需要更好的量化优良性准则。为此,本文用因子分析的基本思想,提出因子分析中降维性、解释性、拟合性更好的量化优良性准则,并证明有目标函数最大化的因子分析模型L是因子分析的一个优良模型。实证表明,因子分析模型L具有优良性,传统因子分析模型的最大似然法、主因子法误差较大。Because of the prevalence of the factor analysis model, common factor solution cant＇t be able to reduce dimensionality or explan the loss of a number of variables. Therefore, the factor analysis needs to find a better model, and the confirmation of a better model needs for a good standard. To this end, there is a basic idea of the factor analysis to find superior standards which have better dimensionality reduction, explanation, and the fitness. It proves that the Factor Analysis Model L that has maximize the objective function is a better model for factor analysis. The actual data has been illustrated that Factor Analysis Model L is better than the maximum likelihood method and the principal factor method.

关 键 词：因子分析模型L 优良模型 实证分析 

分 类 号：F064.1[经济管理—政治经济学]