11 Techniques: PCA - Selected Topics Flashcards
1
Q
What is the modeling quality Ψq?
A
the proportion of explained variance by the first q ≤ p principal components
2
Q
What is the modeling quality Ψq if q = p ?
A
then Ψp = 1
3
Q
What are the charactersitics (mean, covariance) of the normalized principal components matrix Xₛ?
A
x̅ₛ = 0 ^Σ = ^P with Spectral Decomposition ^Γₚ^Λₚ^Γₚᵀ
4
Q
How would you interpret Ψq = x ?
A
With the first q principal components, we can explain x*100% of the data variation
5
Q
Which values are you looking for when conducting principal components analysis?
A
estimated:
- μ, Σ,
- Γ (usually given), Λ (usually given),
- principal component Y
- correlation coefficient ρ
- model quality Ψ