YouTip

NumPy includes a matrix library called numpy.matlib. The functions in this module return matrices, not ndarray objects.

numpy.matlib is specifically designed for creating and manipulating matrix objects. While NumPy's core supports multi-dimensional arrays (ndarray), numpy.matlib provides convenient functions that return matrix types, making it easier for users to perform traditional linear algebra operations.

An matrix is a rectangular array of elements arranged in rows and columns.

The elements in a matrix can be numbers, symbols, or mathematical expressions. Here is a 2-row, 3-column matrix composed of 6 numerical elements:

Transpose Matrix

In NumPy, besides using the numpy.transpose function to swap array dimensions, you can also use the T attribute.

For example, if you have an m-row, n-column matrix, using the T attribute will convert it to an n-row, m-column matrix.

Example

import numpy as np

a = np.arange(12).reshape(3,4)
print('Original array:')
print(a)
print('n')
print('Transposed array:')
print(a.T)

The output is as follows:

Original array:
[
 
 ]

Transposed array:
[
 
 
 ]

matlib.empty()

The matlib.empty() function returns a new matrix. The syntax is:

numpy.matlib.empty(shape, dtype, order)

Parameter Description:

• shape: An integer or tuple of integers defining the shape of the new matrix.
• dtype: Optional, the data type.
• order: 'C' (row-major) or 'F' (column-major).

Example

import numpy.matlib
import numpy as np

print(np.matlib.empty((2,2)))

The output is:

[[-1.49166815e-154 -1.49166815e-154]
 [ 2.17371491e-313  2.52720790e-212]]

numpy.matlib.zeros()

The numpy.matlib.zeros() function creates a matrix filled with zeros.

Example

import numpy.matlib
import numpy as np

print(np.matlib.zeros((2,2)))

The output is:

[[0. 0.]
 [0. 0.]]

numpy.matlib.ones()

The numpy.matlib.ones() function creates a matrix filled with ones.

Example

import numpy.matlib
import numpy as np

print(np.matlib.ones((2,2)))

The output is:

[[1. 1.]
 [1. 1.]]

numpy.matlib.eye()

The numpy.matlib.eye() function returns a matrix with ones on the diagonal and zeros elsewhere.

numpy.matlib.eye(n, M, k, dtype)

Parameter Description:

• n: Number of rows in the returned matrix.
• M: Number of columns in the returned matrix, defaults to n.
• k: Index of the diagonal.
• dtype: Data type.

Example

import numpy.matlib
import numpy as np

print(np.matlib.eye(n = 3, M = 4, k = 0, dtype = float))

The output is:

[[1. 0. 0. 0.]
 [0. 1. 0. 0.]
 [0. 0. 1. 0.]]

numpy.matlib.identity()

The numpy.matlib.identity() function returns an identity matrix of a given size.

An identity matrix is a square matrix where all elements on the main diagonal (from the top-left to the bottom-right) are 1, and all other elements are 0.

Example

import numpy.matlib
import numpy as np

print(np.matlib.identity(5, dtype = float))

The output is:

[[1. 0. 0. 0. 0.]
 [0. 1. 0. 0. 0.]
 [0. 0. 1. 0. 0.]
 [0. 0. 0. 1. 0.]
 [0. 0. 0. 0. 1.]]

numpy.matlib.rand()

The numpy.matlib.rand() function creates a matrix of a given size, filled with random data.

Example

import numpy.matlib
import numpy as np

print(np.matlib.rand(3,3))

The output is:

[[0.23966718 0.16147628 0.14162   ]
 [0.28379085 0.59934741 0.62985825]
 [0.99527238 0.11137883 0.41105367]]

Matrices are always two-dimensional, while ndarray is an n-dimensional array. The two objects are interchangeable.

Example

import numpy.matlib
import numpy as np

i = np.matrix('1,2;3,4')
print(i)

The output is:

[
 ]

Example

import numpy.matlib
import numpy as np

j = np.asarray(i)
print(j)

The output is:

[
 ]

Example

import numpy.matlib
import numpy as np

k = np.asmatrix(j)
print(k)

The output is:

[
 ]

The following functions also exist in the NumPy namespace, but they return matrix objects.

Function Name	Parameter Description	Function Description
matrix(data[, dtype, copy])	data: array-like or string, dtype: data type, copy: whether to copy	Creates a matrix object from an array or string.
asmatrix(data[, dtype])	data: input data, dtype: data type (optional)	Converts input to a matrix object.
bmat(obj[, ldict, gdict])	obj: string or nested sequence, ldict, gdict: namespace dictionaries (optional)	Constructs a matrix from a string, nested sequence, or array.
empty(shape[, dtype, order])	shape: shape, dtype: data type, order: storage order	Creates a matrix with uninitialized elements.
zeros(shape[, dtype, order])	shape: shape, dtype: data type, order: storage order	Creates a matrix filled with zeros.
ones(shape[, dtype, order])	shape: shape, dtype: data type, order: storage order	Creates a matrix filled with ones.
eye(n[, M, k, dtype, order])	n: number of rows, M: number of columns, k: diagonal offset, dtype: data type, order: storage order	Creates a matrix with ones on the diagonal and zeros elsewhere.
identity(n[, dtype])	n: size, dtype: data type (optional)	Creates an identity square matrix.
repmat(a, m, n)	a: array or matrix, m, n: repetition counts	Repeats an array or matrix m times along rows and n times along columns.
rand(*args)	args: specify matrix shape	Generates a random matrix with uniform distribution of the specified shape.
randn(*args)	args: specify matrix shape	Generates a random matrix with standard normal distribution of the specified shape.

Notes

• Although numpy.matlib is convenient, NumPy officially recommends using ndarray and the @ operator for matrix operations, as the matrix type has some limitations and may be deprecated in future versions.
• For new code, it is recommended to prioritize using ndarray to ensure better compatibility and flexibility.