> Linear Algebra Also see NumPy

Python For Data Science

You’ll use the linalg and sparse modules.

Note that scipy.linalg contains and expands on numpy.linalg.

Matrix Functions

SciPy Cheat Sheet

Addition
>>> from scipy import linalg, sparse
>>> np.add(A,D) #Addition

Creating Matrices Subtraction

Learn SciPy online at www.DataCamp.com >>> np.subtract(A,D) #Subtraction

>>> A = np.matrix(np.random.random((2,2)))
Division
>>> B = np.asmatrix(b)
>>> np.divide(A,D) #Division
>>> C = np.mat(np.random.random((10,5)))

>>> D = np.mat([[3,4], [5,6]]) Multiplication

>>> np.multiply(D,A) #Multiplication

SciPy Basic Matrix Routines

>>>
>>>
>>>
np.dot(A,D) #Dot product

np.vdot(A,D) #Vector dot product

np.inner(A,D) #Inner product

Inverse >>> np.outer(A,D) #Outer product

The SciPy library is one of the core packages for

>>> np.tensordot(A,D) #Tensor dot product

>>> A.I #Inverse

>>> np.kron(A,D) #Kronecker product
scientific computing that provides mathematical
>>> linalg.inv(A) #Inverse

algorithms and convenience functions built on the

>>> A.T #Tranpose matrix
Exponential Functions
>>> A.H #Conjugate transposition
>>> linalg.expm(A) #Matrix exponential

NumPy extension of Python. >>> np.trace(A) #Trace >>> linalg.expm2(A) #Matrix exponential (Taylor Series)

Norm >>> linalg.expm3(D) #Matrix exponential (eigenvalue decomposition)

>>> linalg.norm(A) #Frobenius norm

Logarithm Function

> Interacting With NumPy Also see NumPy >>> linalg.norm(A,1) #L1 norm (max column sum)

>>> linalg.norm(A,np.inf) #L inf norm (max row sum)

>>> linalg.logm(A) #Matrix logarithm
Trigonometric Functions
Rank >>> linalg.sinm(D) Matrix sine

>>> import numpy as np

>>> a = np.array([1,2,3])
>>> np.linalg.matrix_rank(C) #Matrix rank >>> linalg.cosm(D) Matrix cosine

>>> b = np.array([(1+5j,2j,3j), (4j,5j,6j)])

>>> linalg.tanm(A) Matrix tangent
Determinant
>>> c = np.array([[(1.5,2,3), (4,5,6)], [(3,2,1), (4,5,6)]]) Hyperbolic Trigonometric Functions
>>> linalg.det(A) #Determinant
>>> linalg.sinhm(D) #Hypberbolic matrix sine

Solving linear problems

Index Tricks >>> linalg.solve(A,b) #Solver for dense matrices

>>> linalg.coshm(D) #Hyperbolic matrix cosine

>>> linalg.tanhm(A) #Hyperbolic matrix tangent

>>> E = np.mat(a).T #Solver for dense matrices

>>> np.mgrid[0:5,0:5] #Create a dense meshgrid

>>> linalg.lstsq(D,E) #Least-squares solution to linear matrix equation Matrix Sign Function
>>> np.ogrid[0:2,0:2] #Create an open meshgrid
>>> np.sigm(A) #Matrix sign function
>>> np.r_[[3,[0]*5,-1:1:10j] #Stack arrays vertically (row-wise)
Generalized inverse
>>> np.c_[b,c] #Create stacked column-wise arrays >>> linalg.pinv(C) #Compute the pseudo-inverse of a matrix (least-squares solver)
Matrix Square Root
>>> linalg.pinv2(C) #Compute the pseudo-inverse of a matrix (SVD) >>> linalg.sqrtm(A) #Matrix square root

Shape Manipulation Arbitrary Functions

Creating Sparse Matrices >>> linalg.funm(A, lambda x: x*x) #Evaluate matrix function
>>> np.transpose(b) #Permute array dimensions

>>>
>>>
b.flatten() #Flatten the array

np.hstack((b,c)) #Stack arrays horizontally (column-wise)

>>>
>>>
F = np.eye(3, k=1) #Create a 2X2 identity matrix

G = np.mat(np.identity(2)) #Create a 2x2 identity matrix

Decompositions
>>> np.vstack((a,b)) #Stack arrays vertically (row-wise)
>>> C[C > 0.5] = 0

>>> np.hsplit(c,2) #Split the array horizontally at the 2nd index

>>> H = sparse.csr_matrix(C) #Compressed Sparse Row matrix
Eigenvalues and Eigenvectors
>>> np.vpslit(d,2) #Split the array vertically at the 2nd index >>> I = sparse.csc_matrix(D) #Compressed Sparse Column matrix
>>> la, v = linalg.eig(A) #Solve ordinary or generalized eigenvalue problem for square matrix

>>> J = sparse.dok_matrix(A) #Dictionary Of Keys matrix

>>> l1, l2 = la #Unpack eigenvalues

>>> E.todense() #Sparse matrix to full matrix

Polynomials >>> sparse.isspmatrix_csc(A) #Identify sparse matrix

>>>
>>>
v[:,0] #First eigenvector

v[:,1] #Second eigenvector

>>> linalg.eigvals(A) #Unpack eigenvalues

>>> from numpy import poly1d

>>> p = poly1d([3,4,5]) #Create a polynomial object Sparse Matrix Routines Singular Value Decomposition
>>> U,s,Vh = linalg.svd(B) #Singular Value Decomposition (SVD)

Inverse >>> M,N = B.shape

Vectorizing Functions >>> sparse.linalg.inv(I) #Inverse >>> Sig = linalg.diagsvd(s,M,N) #Construct sigma matrix in SVD

Norm LU Decomposition
>>> def myfunc(a): if a < 0:
>>> P,L,U = linalg.lu(C) #LU Decomposition
return a*2
>>> sparse.linalg.norm(I) #Norm
else:
Solving linear problems
return a/2

>>> sparse.linalg.spsolve(H,I) #Solver for sparse matrices

>>> np.vectorize(myfunc) #Vectorize functions

Type Handling Sparse Matrix Functions

>>> sparse.linalg.expm(I) #Sparse matrix exponential
>>> np.real(c) #Return the real part of the array elements

>>> np.imag(c) #Return the imaginary part of the array elements

>>>
>>>
np.real_if_close(c,tol=1000) #Return a real array if complex parts close to 0

np.cast['f'](np.pi) #Cast object to a data type Sparse Matrix Decompositions

>>> la, v = sparse.linalg.eigs(F,1) #Eigenvalues and eigenvectors

Other Useful Functions >>> sparse.linalg.svds(H, 2) #SVD

>>> np.angle(b,deg=True) #Return the angle of the complex argument

>>>
>>>
g = np.linspace(0,np.pi,num=5) #Create an array of evenly spaced values(number of samples)

g [3:] += np.pi
> Asking For Help Learn Data Skills Online at
>>>
>>>
np.unwrap(g) #Unwrap

np.logspace(0,10,3) #Create an array of evenly spaced values (log scale)

www.DataCamp.com
>>> np.select([c<4],[c*2]) #Return values from a list of arrays depending on conditions
>>> help(scipy.linalg.diagsvd)

>>> misc.factorial(a) #Factorial

>>> np.info(np.matrix)
>>> misc.comb(10,3,exact=True) #Combine N things taken at k time

>>> misc.central_diff_weights(3) #Weights for Np-point central derivative

>>> misc.derivative(myfunc,1.0) #Find the n-th derivative of a function at a point