Vizualizace a interpretace výsledků PCA

Unsupervised Learning in R

Hank Roark

Senior Data Scientist at Boeing

Biplot

biplot

Scree plot

Biploty v R

# Creating a biplot
pr.iris <- prcomp(x = iris[-5], 
                  scale = FALSE, 
                  center = TRUE)
biplot(pr.iris)

biplot

Scree ploty v R

# Getting proportion of variance for a scree plot
pr.var <- pr.iris$sdev^2
pve <- pr.var / sum(pr.var)

# Plot variance explained for each principal component
plot(pve, xlab = "Principal Component",
       ylab = "Proportion of Variance Explained",
       ylim = c(0, 1), type = "b")

Scree plot

Pojďme procvičovat!

Unsupervised Learning in R