摘要
【目的】克服以往的相关性指标所存在的线性相关性和信息损失的问题,为实际育种工作提供一种新的多性状间的非线性相关分析方法。【方法】采用四阶矩法,提出一种用于描述变量间或向量间相关性程度的广义相关系数。【结果】将该广义相关系数应用于生物性状团的相关分析,可使性状团之间的相关程度介于0与1。【结论】该广义相关系数计算简单,无信息损失,并能度量变量之间的非线性相关性。
[Objective] The study was to overcome the problem of linear correlation and information loss,and provide a new way of non-linear correlation analysis in different traits for breeding operator. [Method] A kind of new generalized correlation coefficient based on the four rank moments was built, which could describe the correlation of different variables and vectors. [Result] It was applied in the analysis of trait group, and got the correlation measurement between 0 and 1. [Conclusion] The calculation method was simple, without loss of information and its correlation was non-linear.
出处
《西北农林科技大学学报(自然科学版)》
CSCD
北大核心
2008年第5期191-195,共5页
Journal of Northwest A&F University(Natural Science Edition)
基金
国家自然科学基金项目(30571072)
关键词
广义相关系数
非线性相关
信息损失
性状团
generalized correlation coefficient
non-linear correlation
information loss
trait group