我试图确定2个不同的群体是否在多种数量性状上表现出相似的趋势。我正在寻找统计措施来做到这一点。
基本设置：
Group 1: N1 members (e.g. each member is a species or gene)
Group 2: N2 members
For each members of group 1/2, I have M distinct measurement values (e.g. gene expression in a specific condition, height, weight)
N1 and N2 does not have to be equal
开始数据：
Matrix 1: A1 = N1 X M matrix (measurements for all members of Group 1)
Matrix 2: A2 = N2 X M matrix (measurements for all members of Group 2)
目标：量化第1组和第2组之间的相似性
Looking for a symmetric measure "F" such that F(A1,A2)=F(A2,A1)="real number"
我知道我可以将矩阵A1和A2改为概率分布，并应用KL散度，但KL散度不对称。
任何建议将不胜感激。