Computation of Span and Basis of Vectors by Using Python (Linear Algebra: Vector Spaces)

Linear Algebra

PROGRAMMES IN PYTHON AND C LANGUAGES

UNIT − 1: VECTOR SPACES

COMPUTATION OF SPAN AND BASIS OF VECTORS BY USING PYTHON

The span of a set of vectors is the set of all possible linear combination of those vectors. A basis for the span is a minimal set of linearly independent vectors that still span the same space. In Python, the NumPy library can be used to perform these operations, particularly by finding the row echelon form of a matrix formed by the vectors.

Here's how to compute the span and a basis for a set of vectors using Python:

Python

import numpy as np

def compute_basis_and_span (vectors):

" " "

computes a basis for the span of a set of vectors.

Args:

  vectors (list of of np.array): A list of 1D NumPy arrays representing the vector.

Returns:

tuple: A tuple containing:

 - list of np.array: A list of vectors forming a basis for the span.

- str: A description of the span.

" " "

if not vectors:

return [], "The span of an empty set of vectors is the zero vector space.

# Stack the vectors as rows to form a matrix

matrix = np.array (vectors)

# Compute the reduced row echelon form (RREF)

#This process helps helps identify linearly independent vectors

#and implicity describes the span.

#Note: NumPy's linalg.matrix_rank can help determine the dimension of the span

#but finding the basis vectors explicitly often involves RREF or similar technical

For a direct basis, we can use QR decomposition or SVD, Debut for simplicity

# and to illustrate the concept of linear independence, we will focus on the rank.

#and that idea of selecting linearly independent vectors

#Find the rank of the matrix, which is the dimension of the span rank = np.linalg.matrix_rank (matrix)

#To find a basis, we can select linearly independent columns from the original

#corresponding to the pivot columns in the RREF.

#A simpler approach for finding a basis from a set of vectors is to remove

#Linear dependent vectors until only a Linearly independent set remains.

#This can be done by iteratively checking Linear independence.

basis_vectors = []

for vec in vectors:

#Try to add the vector to the basis and check if it's Linearly independent

temp_basis_matrix = np.array (basis_vectors + [vec])

if np.linalg.matrix_rank (temp_basis_matrix) > len (basis_vectors):

basis_vectors.append(vec)

#Describe the span based on the dimension

if rank == len (vectors [0]):

span_description = f" The span is R^(len (vectors [0])}

(the entire (len (vector))

elif rank == 0';

span_description = "The span is the zero vector space (origin)."

else:

span_description = f"The span is a {rank)−dimensional subspace of R^(len (vector)

return basis_vectors, span_description}

#Example Usage:

vectors1 = [np.array([1, 0, 0]), np.array([0,1,0]), np.array([0, 0, 1])]

basisl, spanl = compute basis_and_span (vectors1)

print ("vectors 1:")

print ("Basis:", [v.tolist () for V in basis1])

print ("Span: span1)

vectors2 = [np.array([1, 2]), np.array([2, 4])]

basis2, span 2 = compute_basis_and_span (vectors2)

print ("\nvectors 2")

print ("Basis: ",[v.tolist() for V in basis2])

print ("Span:", span2)

vectors3 = [np.array([1, 0, 1]), np.array([0, 1, 1,]), np.array([1, 1, 2])]

basis3, span3 = compute_basis_and_span (vectors3)

print ("\nvectors 3:")