读入数据

pca = read.table("714_1.txt",header = T)
head(pca)

image.png

去除第一列

pca1 = pca[,2:9]
head(pca1)

image.png

PCA分析

pcadev=princomp(pca1,cor=T)
summary(pcadev,loadings = T)
comp1=pcadev$loadings[,1]
comp2=pcadev$loadings[,2]
comp3=pcadev$loadings[,3]

image.png

Propotion of Variance就是我们要的主成分，Comp.1和Comp.2分别为第一和第二主成分，解释度分别为80.72%和10.34%。推测前三个主成分即可解释90%以上差异。因此选择前三组即可。

出图

因为有八组数据所有需要提供八种颜色，每种颜色从前往后以此对应。

plot3d(comp1,comp2,comp3,col=c("red","gray0","blue","cyan","darkblue","green","darkgreen","lightpink"),size = 10,xlab="PC1",ylab="PC2",zlab="PC3")
rgl.spheres(comp1,comp2,comp3,r=0.03,col=c("red","gray0","blue","cyan","darkblue","green","darkgreen","lightpink"))