中国汉语方言多样性与生物多样性的结构方程建模分析  

Structural Equation Modeling of Linguistic Diversity and Biodiversity in China

作　　者：魏琰 王春霞 邹晨晨 WEI Yan;WANG Chun-xia;ZOU Chen-chen(School of Mathematics and Statistics,Qingdao University,Qingdao 266071,China)

机构地区：[1]青岛大学数学与统计学院,青岛266071

出　　处：《青岛大学学报（自然科学版）》2023年第4期8-12,共5页Journal of Qingdao University(Natural Science Edition)

摘　　要：为研究中国汉语方言多样性与生物多样性间相关性,建模分析了生物多样性、汉语方言多样性和公因子三者之间的关系。利用结构方程模型,综合分析去除公因子影响后,由多指标构建的生物多样性和汉语方言多样性这两个潜变量之间的关系,并采用Lasso、MCP惩罚参数和AMOS基于指定搜索方法选择模型。各省公因子得分与若干宏观指标的Spearman相关系数显示,结构方程所提炼公因子与各省面积、园林草湿地覆盖率、及森林覆盖率呈显著正相关,与各省人口密度呈显著负相关,与各省光照条件、GDP无显著相关性。研究结果表明,去除面积、植被公因子后,生物多样性与语言多样性分别呈微弱的负相关,统计学上不显著。Based on 30 provinces′statistics in China,the Structural Equation Model(SEM)was established to analyze the relationship between the diversity of Chinese dialects and biodiversity after removing common factors.There was a benefit method to optimize the SEM model using model selection such as Lasso,MCP penalty parameter method,and Amos based on specified search method.It was found that common factors are related to the vegetation coverage and area of various provinces in China.The research results show that there is a weak negative correlation between the two after removing the common factors such as area and vegetation.

关 键 词：汉语方言多样性 生物多样性 结构方程模型 模型选择 语言生态学 

分 类 号：O213[理学—概率论与数理统计]