Introduction to PCA

Unsupervised Learning in R

Hank Roark

Senior Data Scientist at Boeing

Two methods of clustering

Two methods of clustering - finding groups of homogeneous items
Next up, dimensionality reduction
- Find structure in features
- Aid in visualization

Dimensionality reduction

A popular method is principal component analysis (PCA)
Three goals when finding lower dimensional representation of features:
- Find linear combination of variables to create principal components
- Maintain most variance in the data
- Principal components are uncorrelated (i.e., orthogonal to each other)

PCA intuition

scatter plot

PCA intuition

regression line

PCA intuition

component scores projection

Visualization of high dimensional data

different number of dimensions for data

Visualization

PCA on iris dataset

PCA in R

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

summary(pr.iris)

Importance of components:
                          PC1     PC2    PC3     PC4
Standard deviation     2.0563 0.49262 0.2797 0.15439
Proportion of Variance 0.9246 0.05307 0.0171 0.00521
Cumulative Proportion  0.9246 0.97769 0.9948 1.00000

Let's practice!

Unsupervised Learning in R