论文部分内容阅读
如果通过若干个体测定的性状不只一个,并且这些性状之间的关系必须同时加以研究的话,我们就要用到多元分析方法。在研究个体间存在的相互关系时,常常使用一个简单的几何模型。一个需要测定 p 个性状的个体,可以表示成 P 维空间中的一个点;它在空间直角座标系中第 j 个座标轴上的座标,由第 j 个性状值给出。当 P>2时,这种表示方法很难给出直观的解释。本文的目的就是,在多元情形下综合一类比较成熟的有效地减少空间维数的方法,并对原始数据保持足够详细的描述。
If there are more than one trait measured by several individuals, and the relationships between these traits must be studied at the same time, then we need to use multivariate analysis. When studying the interrelationships between individuals, a simple geometric model is often used. An individual who needs to determine p traits can be expressed as a point in the P-dimensional space. The coordinate of the j-th coordinate in the space Cartesian coordinate system is given by the jth trait value. When P> 2, this method is difficult to give an intuitive explanation. The purpose of this article is to synthesize a more sophisticated and efficient method of reducing the number of spatial dimensions in a multitude of contexts and to maintain a sufficiently detailed description of the original data.