Matrix Operations.pdf

(189 KB) Pobierz
179605034 UNPDF
Matrix Operations
Yu Jiangsheng
Institute of Computational Linguistics
Peking University
September 26, 2002
Top i cs
1. Matrix Multiplication
2. Solving System of Linear System
3. Inverting Matrixes
4. Symmetric Positive-definitive Matrixes and
Least-squares Approximation
5. Winograd Theorem and AHU Theorem
1
Matrix
Amatrix A is, in fact, a sequence of arrays:
a 11
a 12
···
a 1 n
= a ij m×n
a 21
a 22
···
a 2 n
A =
.
a m 1
a m 2
···
a mn
Note From the viewpoint of transformation,
Ax describes the rotation of x around 0 and
linear stretching.
2
179605034.001.png 179605034.002.png 179605034.003.png 179605034.004.png
Singular Matrix
linearly independent row rank and column
rank rank full rank existence of in-
verse (nonsingular)
Definition 1 A null vector for a matrix A m×n
is a nonzero vector x such that Ax =0.
Homework 1 Amatrix A has full column rank
iff it has no null vector.
Homework 2 A square matrix is singular iff
it has null vector.
3
Determinant of Matrix
1) matrix A [ ij ] obtained
by deleting the i th row and the j th column
of A .
Definition 3 The determinant of A n×n is de-
j =1 ( 1) 1+ j a 1 j det( A [1 j ] )f n> 1
if n =1
det( A )=
(1)
Property 1 A n×n is singular iff det( A )=0.
4
Definition 2 The ij th minor of matrix A n×n
is the ( n
1) × ( n
fined recursively in terms of minors by
a 11
n
Zgłoś jeśli naruszono regulamin