定量古生物学中的常用Q型多变量线性回归方法例析  

Q-MODE MULTIVARIATE REGRESSION ANALYSES IN QUANTITATIVE PALEONTOLOGY AND EXAMPLES

作　　者：宋思存 张子航 史宇坤[1] SONG Sicun;ZHANG Zihang;SHI Yukun(Centre for Research and Education on Biological Evolution and Environment&Frontiers Science Center for Critical Earth Material Cycling,Nanjing University,Nanjing 210023)

机构地区：[1]南京大学生物演化与环境科教融合中心,关键地球物质循环前沿科学中心(南京大学),南京210023

出　　处：《微体古生物学报》2022年第2期190-199,共10页Acta Micropalaeontologica Sinica

基　　金：由国家自然科学基金项目(批准号:41772017,91955201);中央高校基本科研业务费专项资金(0206-14380121);关键地球物质循环前沿科学中心科研基金(JBGS2101)联合资助。

摘　　要：提要Q型多变量线性回归方法是定量古生物学研究中经常使用的一组方法。不同于把研究重心放在变量或变量组之间关系的R型回归方法,Q型回归方法对样本进行整体分析,通过构建样本之间的相似性系数矩阵或者距离矩阵,并进行分解与变换,进而用可视化的方法展示样本的整体相似性关系。本文将介绍常用的Q型多变量线性回归方法,包括主坐标分析(Principal coordinate analysis)、Q型聚类分析(Q-Cluster analysis)、Q型因子分析(Q-Factor analysis),以及综合了R型和Q型的对应分析(Correspondence analysis),和它们在计算中的注意事项与适用场景。本文选取瓜迪亚纳河入海口大陆架地区底栖有孔虫分布特征研究作为实例,分别用上述方法进行分析,以解释它们在定量古生物学中的具体应用。Q-mode multivariate regression analyses are useful statistical methods in the quantitative paleontology.Different from the R-mode analyses which focus on the configuration of variables, Q-mode analyses pay attention to the samples as a whole. The decomposition is processed to the distance or similarity matrices of the original fossil data,and then transformations might be needed for various purposes. The acquired results can be plotted and illustrate the relationship of all the samples. This paper introduces the commonly used Q-mode multivariate regression methods,including principal coordinate analysis(PCoA), Q-cluster analysis, and Q-factor analysis, as well as a special analysis called CA, correspondence analysis, which integrate the characteristics of the R-mode and Q-mode analyses. The benthic foraminiferal samples in the continental shelf area of Guadiana River Estuary is taken as an example to be analyzed by the above-mentioned methods and explain the application of Q-mode multivariate regression analyses in the quantitative paleontology.

关 键 词：Q型多变量线性回归 主坐标分析 Q型聚类分析 Q型因子分析 对应分析 

分 类 号：Q91[天文地球—古生物学与地层学]