Chapter 2

Question 1

What is important about the transformed cov matrix Sy?(2)

Answer 1

It is diagonal, the transformed variables are uncorrelated with one another-note that Y is a linear transformation of X (Y=AX^T)
and that y1 has largest sample variance, y2 second largest etc.
A is a rotation matrix and the jth PC axis is defined yj=aj^Tx for x the original axes.
y1 is the first PC.

Question 2

Derivation of PCs 2.2.2

Answer 2

Need to know this!!