关于因子模型L的讨论  

Discussion on Factor Model L

作　　者：李选举 Li Xuanju(School of Digital Economics and Trade,Guangzhou Huashang College,Guangzhou 511300,China;College of Economics,Shenzhen University,Shenzhen Guangdong 518060,China)

机构地区：[1]广州华商学院数字经贸学院,广州511300 [2]深圳大学经济学院,广东深圳518060

出　　处：《统计与决策》2024年第24期36-41,共6页Statistics & Decision

基　　金：广东省重点建设学科科研能力提升项目(2022ZDJS127)。

摘　　要：文献[1—3]认为经典因子模型存在诸多缺陷,因此提出了因子模型L,认为因子模型L改进了经典因子模型,得到了最优解,且具有诸多优良性质。文章对此提出了质疑,并通过论证得到了以下结论:(1)因子模型L关于假设条件的“放宽”和“极大值”条件的增加没有意义;因子模型L与经典模型主成分分析法的载荷阵相同。因此,因子模型L对经典因子模型没有改进。(2)因子模型L的载荷阵不是最优解,并且因子模型L的误差有遗漏。(3)因子模型L不能沿用经典模型的因子得分公式,但是在计算时仍然沿用经典模型的计算公式,其理论模型与实际计算自相矛盾。Literature[1—3]considers that the classical factor model has many defects,so Factor Model L is proposed,and points out that Factor Model L improves the classical factor model,obtains the optimal solution,and has many excellent properties.This paper raises the question about this,and obtains the following conclusion through argumentation:(1)Factor Model L does not make sense with respect to the“relaxation”of the assumptions and the increase of the“maximum”conditions;Factor Model L is the same as the load matrix of the classical model principal component analysis.Therefore,Factor Model L does not improve the classical factor model.(2)The load matrix of Factor Model L is not the optimal solution,and the error of Factor Model L is missing.(3)Factor Model L cannot follow the formula of factor score of the classical model,but it still follows the formula of the classical model in time of calculation,and its model assumptions contradict the actual calculation.

关 键 词：经典因子模型 因子模型L 主成分分析法 

分 类 号：O212.1[理学—概率论与数理统计]