1, the scalar (Scalar), vector (Vector), matrix (Matrix), tensor (tensor).
2, some of the concepts of the matrix: the main diagonal (main diagonal), matrix (identity matrix), the inverse matrix (matrix inversion), the diagonal matrix (diagonal matrix), symmetric matrix (symmetric matrix), orthogonal moments
阵（orthogonal matrix）。
3, for some operations and matrix operations: transposition (TRANSPOSE), a scalar matrix adding a scalar matrix multiplication, broadcast (Broadcast) (the sum of each row of the matrix and vector), a vector matrix multiplication, inversion.
4, the matrix product / standard product (matrix product): C = AB
Element corresponding to the product / Hadamard product: C = A⊙B
5, the linear correlation linearly independent.
6, norm (norm): with a common norm ① L . 1 norm: || X || . 1 = [Sigma | X I | ②L 2 -norm (Euclidean norm) ③ L ∞ norm (maximum norm ): || X || ∞ = max | X I |
When → measure the size of the matrix, common depth study Frobenius norm: || A || F. = ([Sigma ( A I, J ) 2 ) 1/2
7, the dot product of the vector may be expressed as: X T Y = || X || 2 || Y || 2 cos [theta].
8, wherein decomposition (eigendecomposition): the matrix into a set of eigenvalues and eigenvectors. Have any of a real symmetric matrix decomposition characteristics, but may not be unique . Depending on the values of the positive and negative characteristics, it may be referred to as a positive definite matrix
(Positive definite), semi definite (positive semidefinite), negative definite (negative definite), semi-definite matrix minus (negative semidefinite).
9, singular value decomposition (singular value decomposition, SVD): the matrix into singular vectors and singular values. Each solid matrix has a singular value decomposition .
10, Moore-Penrose pseudo-inverse
11, the operational track Tr (A): Returns the matrix and the diagonal elements. Scalar operations are trace itself.
12, the determinant det (A)