Chapter 4 & 5

PCA

Subtract the mean

Compute the singular value decomposition (SVD)

Plot data in the transformed

$\tilde{x}=x-\bar{x}$

$b^T=\tilde{x}^T[v_1, v_2]$

$x' = b[v_1,v_2]$

Simlarity / Dissimilarity measures

Similarity s(x,y) Often between 0 and 1. Higher are more sililar

Dissimilarity d(x,y) Greater than 0. Lower means more similar

Dissimiilarity measures One-norm Euclidian Max-norm p-distance

If you want to be sensitive to outliers use max-norm distance

Similarity measures