刺槐数量性状的主成分分析及分类

刺槐20个数量性状主成分分析研究结果表明:前5个主成分累计贡献率达85.93％,数量性状可分为4大类群,类群内部因子是正相关,类群之间因子是负相关或不相关。根据主成分因子负荷量及其主要指标组合,至少可看出5个因子对立木的作用。采用相关系数矩阵进行分类,优于用协方差矩阵进行分类。

Major Components Analysis and Classification of Quantitative Characters of Robinia pseudoacacia

Describes the major components analysis of 20 quantitative characters of black locust. The percentage of characteristic value of first 5 major components sumed up to 85.93%. All 20 quantitative characters were classified into4 groups. The factors within each group correlated positively to each other but factors among groups showed negative or no correlation. According to the major component factors load and their main index combinations, it is obvious that at least 5 factors are effective on the living tree. Classification of quantitative factors by correlation coefficient matrix is better than that by covariance matrix.