Answer on Question #44695 – Math - Linear Algebra
Give some detail explanation on Pseudo inverse matrix??
Solution
Pseudoinverse matrix – is a generalization of the inverse matrix in mathematics, particularly in linear algebra.
Pseudoinverse satisfies the following criteria:
- ( is not necessarily equal to the identity matrix);
- ;
- ;
- – also Hermitian matrix);
where – Hermitian-conjugate matrix to the matrix .
Calculation
With schedule
Let – rank matrix size times . Then can be represented as , where – matrix of size times , – matrix of size times . Then
or
where – a smaller matrix of size times .
Using QR decomposition
A matrix represented as , where – unitary matrix, , and – upper triangular matrix. Then
Properties
- Pseudoinverse matrix always exists and is unique.
- Pseudoinverse matrix is equal to zero its transposition.
- Pseudoinverse is reversible to himself:
- Pseudoinverse commutes with transposition, Hermitian coupling and coupling:
- (A ^ T) ^ + = (A ^ +) ^ T, \ qquad (\ overline {A}) ^ + = \ overline {A ^ +}, \ qquad (A ^ *) ^ + = (A ^ +) ^ *.
-Pilot matrix equals its rank to pseudoinverse:
- rank \ A ^ + = rank \ A
- Pseudoinverse matrix product of A by a scalar \ alpha is the product of the matrix A ^ + on inverse number \ alpha ^ {- 1}:
(\ alpha A) ^ + = \ alpha ^ {- 1} \; A ^ +, \ quad \ forall \ alpha \ ne 0.
-If already known matrix (A ^ * A) ^ + or matrix (AA ^ *) ^ + , they can be used to calculate A ^ +:
- A ^ + = (A ^ * A) ^ + \ ; A ^ *
- A ^ + = A ^ * \ ; (A A ^ *) ^ +.
-Matrix \ A ^ + A, \ AA ^ + - is the orthogonal projection-matrices.
-If the matrix \ A_i formed from the matrix \ A by inserting another zero row / column in the i-th position, then A_i ^ + will be created with \ A ^ + by adding a zero column / row in the i-th position.
www.AssignmentExpert.com