PCA

Created
Created
2023 May 4 4:50
Editor
Creator
Creator
Seonglae ChoSeonglae Cho
Edited
Edited
2024 Oct 28 14:2
Refs
Refs
SVD
LDA

Principal component analysis

minimize projection error, maximize projection variance (more original information)

PCA is a special shallow case of
AutoEncoder
equivalent to calculating the
Eigenvector
of the data
Covariance Matrix
corresponding to the largest
Eigenvalue
Also PCA solution means choosing (D − M)-smallest eigenvalues, M-largest eigenvalues
서로 연관 가능성이 있는 고차원 공간의 표본들을 선형 연관성이 없는 저차원 공간의 표본으로 변환하기 위해 직교 변환을 사용
Feature Extraction, Data Compression, Data Visualization
If high-dimensional data, we reduce computation O() to O() using
Kernel PCA
notion image
notion image
PCA Notion
 
 
 
 
 
 

Recommendations