Lecture Note 1 – Linear Algebra Shuaiqiang Wang Department of CS & IS University of Jyväskylä
Vectors Vector Facts: A vector is a point in a coordinate system! A vector has length and direction! (2,1) The length is The direction is
Subtract/Add Vectors (3,3) (1,2) (2,1)
Product of Vectors
Matrix Facts: A matrix is a collection of numbers! A matrix is a collection of vectors! A matrix is a mapping system of vectors!
Special Matrix
Transpose of Matrix
Subtract/Add Matrix
Product of Matrix
Inverse of Matrix
Mapping of Vectors Mapping SystemOriginal VectorMapped Vector How does it work? A matrix is a mapping system of vectors!
Simple Demonstration We use the 2-D system to demonstrate the mapping process Easy to calculate its eigenvalues and eigenvectors
Factorization of Square Matrix Square matrix Matrix composed of eigenvectors Diagonal matrix composed of eigenvalues Transpose of matrix For example:
Example Since Points (vectors) in the blue area are mapped into the red area! The directions of the mapping (in dashed lines) are decided by the directions of the eigenvectors! The strength of the mapping in two directions can be decided by the values of the eigenvalues!
Non-Square Matrix Square Matrix They share same eigenvalues! The vectors in P or Q are unit ones and orthogonal to each other!
Singular Value Decomposition (SVD)
Principal Component Analysis
We can keep the main mapping directions while remove some trivial directions:
Example Application in IR
SVD Application
Any Question?