Operations - Numpy
January 6, 2021
[1]: import numpy as np
[2]: # Define Two 3x3 Matrices
A = np.array([[2, 6, 4], [9, 1, 3], [8, 5, 7]])
B = np.array([[6, 8, 1], [7, 9, 5], [8, 2, 4]])
[3]: print('Mathematical Operations Using Numpy: ')
Mathematical Operations Using Numpy:
[4]: # Addition
A + 1
[4]: array([[ 3,
[10,
[ 9,
7,
2,
6,
5],
4],
8]])
[5]: A + B + 6
[5]: array([[14, 20, 11],
[22, 16, 14],
[22, 13, 17]])
[6]: # Substraction
A - 2
[6]: array([[ 0, 4,
[ 7, -1,
[ 6, 3,
2],
1],
5]])
[7]: A - B - 3
[7]: array([[ -7, -5,
[ -1, -11,
[ -3,
0,
0],
-5],
0]])
1
[8]: # Multiply By a Constant
6 * A
[8]: array([[12, 36, 24],
[54, 6, 18],
[48, 30, 42]])
[9]: # One-to-One Multiplication
A * B
[9]: array([[12, 48, 4],
[63, 9, 15],
[64, 10, 28]])
[10]: # Devide By a Constant
A / 3
[10]: array([[0.66666667, 2.
, 1.33333333],
[3.
, 0.33333333, 1.
],
[2.66666667, 1.66666667, 2.33333333]])
[12]: # One-to-One Division
A / B
[12]: array([[0.33333333, 0.75
, 4.
[1.28571429, 0.11111111, 0.6
[1.
, 2.5
, 1.75
],
],
]])
[13]: # Modulus By a Constant (remainder by division)
A % 6
[13]: array([[2, 0, 4],
[3, 1, 3],
[2, 5, 1]], dtype=int32)
[14]: # One-to-One Modulus (remainder by division)
A % B
[14]: array([[2, 6, 0],
[2, 1, 3],
[0, 1, 3]], dtype=int32)
[16]: # Exponent As Array Eelements
2 ** A
[16]: array([[ 4,
[512,
64,
2,
16],
8],
2
[256,
32, 128]], dtype=int32)
[18]: # One-to-One Exponentiation (Array Eelement Raised to Corresponding Array␣
,→Eelement)
A ** B
[18]: array([[
64,
[ 4782969,
[16777216,
1679616,
1,
25,
4],
243],
2401]], dtype=int32)
[23]: # Mixture of Operations":
2 ** (A + B) * (A + B) / A + 1
[23]: array([[1.02500000e+03, 3.82303333e+04, 4.10000000e+01],
[1.16509444e+05, 1.02410000e+04, 6.83666667e+02],
[1.31073000e+05, 1.80200000e+02, 3.21928571e+03]])
[24]: # Since We gone through One-to-One Multiplication:
# Now is the Time To Take a Look At Mathematical Matrix Multiplication
# In This Case We Multiply Two Matrix (No Scalars All Elements Should be␣
,→Matrices)
A.dot(B)
[24]: array([[ 86, 78,
[ 85, 87,
[139, 123,
48],
26],
61]])
[26]: # Another Handy Way to do So is by Using the (At) Operator
[email protected]
[26]: array([[ 86, 78,
[ 85, 87,
[139, 123,
48],
26],
61]])
[29]: # Element Comparisons
A == B
[29]: array([[False, False, False],
[False, False, False],
[ True, False, False]])
[30]: A != B
[30]: array([[ True,
[ True,
[False,
True,
True,
True,
True],
True],
True]])
3
[31]: # In case we needed the Boolean Answer:
np.array_equal(A, B)
[31]: False
[32]: np.array_equal(A, A)
[32]: True
[36]: # Logical Operations: Or & And
# Lets Create Two arrays of Bool Values
b1 = np.array([0, 1, 0, 1, 1, 1])
b2 = np.array([1, 1, 0, 1, 0, 1])
[38]: np.logical_or(b1, b2)
[38]: array([ True,
True, False,
True,
True,
True])
True, False,
True])
[39]: np.logical_and(b1, b2)
[39]: array([False,
True, False,
[40]: # Mapping Numpy Math-Built-in Functions all over elements
np.sin(A)
[40]: array([[ 0.90929743, -0.2794155 , -0.7568025 ],
[ 0.41211849, 0.84147098, 0.14112001],
[ 0.98935825, -0.95892427, 0.6569866 ]])
[41]: np.cos(B)
[41]: array([[ 0.96017029, -0.14550003, 0.54030231],
[ 0.75390225, -0.91113026, 0.28366219],
[-0.14550003, -0.41614684, -0.65364362]])
[42]: np.tan(A + B)
[42]: array([[
[
[
-6.79971146,
0.30063224,
0.30063224,
7.24460662,
-3.38051501],
0.64836083,
-6.79971146],
0.87144798, -225.95084645]])
[56]: np.sqrt(A + B)
[56]: array([[2.82842712, 3.74165739, 2.23606798],
[4.
, 3.16227766, 2.82842712],
[4.
, 2.64575131, 3.31662479]])
4
[43]: np.floor(A / 6)
[43]: array([[0., 1., 0.],
[1., 0., 0.],
[1., 0., 1.]])
[44]: np.ceil(9 / B)
[44]: array([[2., 2., 9.],
[2., 1., 2.],
[2., 5., 3.]])
[47]: # Median , Mean, Average, Standard Deviation & Variance
a = np.array([1, 2, 6, 9, 7, 2, 8, 2, 6])
print(f'Median
of a is: {np.median(a)}')
print(f'Mean
of a is: {np.mean(a)}')
print(f'Average
of a is: {np.average(a)}')
print(f'Standard Dev of a is: {np.std(a)}')
print(f'Variance
of a is: {np.var(a)}')
Median
of
Mean
of
Average
of
Standard Dev of
Variance
of
a
a
a
a
a
is:
is:
is:
is:
is:
6.0
4.777777777777778
4.777777777777778
2.8588178511708016
8.172839506172838
[48]: # Log & Exp
np.log(A)
[48]: array([[0.69314718, 1.79175947, 1.38629436],
[2.19722458, 0.
, 1.09861229],
[2.07944154, 1.60943791, 1.94591015]])
[49]: np.exp(B)
[49]: array([[4.03428793e+02, 2.98095799e+03, 2.71828183e+00],
[1.09663316e+03, 8.10308393e+03, 1.48413159e+02],
[2.98095799e+03, 7.38905610e+00, 5.45981500e+01]])
[53]: # Array Transposition(The Flip of the Matrix Over the Main Diagonal)
np.transpose(A)
[53]: array([[2, 9, 8],
[6, 1, 5],
[4, 3, 7]])
5
[54]: # Or the Handy Way
A.T
[54]: array([[2, 9, 8],
[6, 1, 5],
[4, 3, 7]])
[55]: B.T
[55]: array([[6, 7, 8],
[8, 9, 2],
[1, 5, 4]])
[ ]:
6