论文部分内容阅读
n个群体的p维环境因子主成分分析(PCA),Z=U’X,恢复原始数据线性表达成。依据λ值,取前1—Υ个主成分,1≤Υ≤p,将p维环境因子降维成1元向量,建立前Υ个主成分的生态梯度轴[EGA(PC_r)],估算误差为d_(i)。对白榆全分布区20个群体的样点6维环境因子PCA求算的EGA(PC_r),取r=1,2,3,建立3个生态梯度轴,累计贡献率依次为63.8％,84.3％和96.0％,都能较好的代表环境因子。EGA(PC_r)在揭示群体7个性状的梯度变异中得到验证,并能用于判别种群变异模式。
The principal component analysis (PCA) of p-dimensional environmental factors of n groups, Z = U’X, restored the original data linearly expressed as. According to the value of λ, taking the first 1-Y principal components and 1≤Y≤p, the p-dimensional environmental factors are reduced to 1-element vectors and the ecological gradient axis [EGA (PC_r)] of the first Y principal components is established to estimate the error For d_ (i). The EGA (PC_r) calculated from the PCA of 6-dimensional environmental factor PCA for 20 populations in the whole distribution area of Elm Ulmus, taking r = 1, 2 and 3, established three ecological gradient axes with cumulative contribution rates of 63.8% and 84.3% And 96.0%, all can better represent the environmental factor. EGA (PC_r) was validated in revealing the gradient variation of seven traits in population and could be used to discriminate population variation patterns.