| 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. |
| 关键词: 广义相关系数 非线性相关 信息损失 性状团 |
| DOI： |
| 分类号: |
| 基金项目:国家自然科学基金项目(30571072) |
|
| Non-linear generalized correlation coefficient ofbiology trait |
|
|
| Abstract: |
| 【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. |
| Key words: generalized correlation coefficient non-linear correlation information loss trait group |
