Scribd
Upload
89 views56 pages

Slot06 Numpy1

The document discusses NumPy, which is a Python library for working with arrays. It provides fast array-oriented operations and flexible broadcasting capabilities. The document covers NumPy array basics like creation, shapes, types, indexing and slicing. It also discusses various numerical operations on arrays like arithmetic, logical and matrix operations. Several exercises on manipulating arrays are presented at the end.

Uploaded by

Eric Nam
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
89 views56 pages

Slot06 Numpy1

The document discusses NumPy, which is a Python library for working with arrays. It provides fast array-oriented operations and flexible broadcasting capabilities. The document covers NumPy array basics like creation, shapes, types, indexing and slicing. It also discusses various numerical operations on arrays like arithmetic, logical and matrix operations. Several exercises on manipulating arrays are presented at the end.

Uploaded by

Eric Nam
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 56

University of Science – VNU-HCM

Faculty of Information Science


Department of Computer Science
MTH083 - Advanced Programming for Artificial Intelligence

Slot 06-
Numpy – Part 01
Advisor:
Dr. Nguyễn Tiến Huy
Dr. Lê Thanh Tùng

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 1
Numpy
▪ NumPy is a Python library used for working with arrays
▪ One of the most important foundational packages for numerical
computing in Python.
▪ Most computational packages providing scientific functionality
use NumPy’s array objects as the lingua franca for data
exchange.
▪ Providing fast array-oriented operations and flexible
broadcasting capabilities
Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 2
Array
▪ NumPy is used to work with arrays. The array object in NumPy is
called ndarray
▪ Create a NumPy ndarray object by using the array() function

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 3
Array

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 4
Dimensions in Arrays
▪ A dimension in arrays is one level of array depth (nested arrays)
▪ 0-D array or Scalars, are the elements in an array
▪ 1-D array has 0-D arrays as its elements, called uni-dimensional
▪ 2-D array has 1-D arrays as its elements, called 2-d matrix
▪ 3-D array has 2-D matrices as its elements

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 5
Dimensions in Arrays

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 6
Shape & Type of ndarray
▪ Attribute: ndarray.shape – Tuple of array dimensions

▪ Attribute: ndarray.dtype – Data-type of the array’s elements

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 7
Create Array

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 8
dtype

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 9
Type & Type cast

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 10
numpy.arange()
▪ arange([start,] stop[, step,][, dtype])

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 11
numpy.arange()

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 12
numpy.random.randn
▪ random.randn(d0, d1, ..., dn): Return a sample (or samples) from the
“standard normal” distribution.

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 13
numpy.random.randn

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 14
Numerical operations on arrays
▪ Elementwise operations:
▪ With scalars:

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 15
Numerical operations on arrays
▪ Elementwise operations: With array (same dimension):

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 16
Numerical operations on arrays
▪ Matrix multiplication

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 17
Numerical operations on arrays

Comparisons:

Array-wise
Comparisons:

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 18
Numerical operations on arrays
▪ Logical operations

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 19
List vs numpy array in Math operation

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 20
Indexing & Slicing

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 21
Indexing & Slicing

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 22
Indexing & Slicing

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 23
Indexing & Slicing

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 24
Indexing & Slicing

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 25
Copies & Views in ndarray

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 26
Copies & Views in ndarray

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 27
Boolean indexing
▪ Numpy arrays support a feature called conditional selection to generate
a new array of boolean values that state whether each element within the
array satisfies a particular if statement.

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 28
Boolean indexing

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 29
Boolean indexing

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 30
Fancy Index
▪ Besides using indexing & slicing, NumPy provides you with a convenient
way to index an array called fancy indexing.
▪ Fancy indexing allows you to index a numpy array using the following:
▪ Another numpy array
▪ A Python list
▪ A sequence of integers

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 31
Fancy Index

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 32
Fancy Index

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 33
Reshape
▪ Reshaping means changing the shape of an array

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 34
Reshape
▪ Reshaping returns the view of the original array

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 35
Reshape: Unknown Dimension
▪ Pass -1 as the value, and NumPy will calculate this number for you

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 36
Exercise
▪ Ex1: Given an integer matrix of NxM, set negative values in the
matrix to 0

▪ Ex2: Given an integer matrix of NxM, set negative values in


even rows to 0

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 37
Exercise
def setZero4Negative_Index(arr):
shape = arr.shape
for row in range(shape[0]):
for col in range(shape[1]):
if (arr[row][col] < 0):
arr[row][col] = 0

arr1 = np.array([[1, 2, -3], [-1, -2, 3]])


setZero4Negative_Index(arr1)
print(arr1)
Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 38
Exercise
def setZero4Negative(arr):
arr[arr < 0] = 0

arr2 = np.array([[1, 2, -3], [-1, -2, 3]])


setZero4Negative(arr2)
print(arr2)

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 39
Exercise
def setZero4Negative_Index(arr):
shape = arr.shape
for row in range(shape[0]):
for col in range(shape[1]):
if (row % 2 == 0):
if (arr[row][col] < 0):
arr[row][col] = 0

arr1 = np.array([[1, 2, -3], [-1, -2, 3], [-3, -2, 4]])


setZero4Negative_Index(arr1)
print(arr1)

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 40
Exercise
def setZero4Negative(arr):
index = arr < 0
for i in range(len(arr)): # range(arr.shape[0])
if (i % 2 == 0): arr[i][index[i]] = 0

arr2 = np.array([[1, 2, -3], [-1, -2, 3], [-3, -2, 4]])


setZero4Negative(arr2)
print(arr2)

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 41
Exercise
arr3 = np.array([[1, 2, -3], [-1, -2, 3], [-3, -2, 4]])
arrEvenRow = arr3[::2]
arrEvenRow[arrEvenRow < 0] = 0
print(arr3)

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 42
Exercise
Ex3: Create a checkboard matrix NxN as follows:

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 43
Exercise
def createMatrix(n: int):
arr = np.zeros((n, n), dtype = np.int64)
mark = 0
for row in range(n):
for col in range(n):
arr[row][col] = mark
mark = not(mark)
mark = not(mark)
return arr

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 44
Exercise
def createMatrixVer2(n: int):
arr = np.zeros((n, n), dtype = np.int64)
arr[1::2, ::2] = 1
arr[::2, 1::2] = 1
return arr

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 45
Exercise
Ex4: Create a matrix NxM with 1 in borders and 0 for inside:

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 46
Exercise
def createMatrix(n: int, m: int):
arr = np.zeros((n, m), dtype = np.int64)
for row in range(n):
for col in range(n):
if (row == 0 or row == n-1):
arr[row][col] = 1
if (col == 0 or col == m - 1):
arr[row][col] = 1

return arr

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 47
Exercise
def createMatrixVer2(n: int, m: int):
arr = np.zeros((n, m), dtype = np.int64)
arr[0,:] = 1
arr[n-1,:] = 1
arr[:, 0] = 1
arr[:, m-1] = 1
return arr

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 48
Exercise
def createMatrixVer3(n: int, m: int):
arr = np.ones((n, m), dtype = np.int64)
arr[1:-1, 1:-1] = 0
return arr

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 49
Exercise
Ex5: Create a NxN matrix as follows:
N=4

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 50
Exercise
def createMatrix(n: int):
content = list(range(n*n))
arr = np.array(content)
arr = arr.reshape(n, -1)
for row in range(n):
for col in range(n):
if (col == row):
arr[row][col] = 100
return arr

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 51
Exercise
def createMatrixVer2(n: int):
arr = np.arange(n*n)
arr[0::5] = 100
return arr.reshape(n, -1)

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 52
Exercise
Ex6: Write a function to multiply two matrices NxN without using
np.dot()

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 53
Exercise
def matrix_multiply(a, b):
n = len(a)
c = np.array([[0 for i in range(n)] for j in range(n)])
for i in range(n):
for j in range(n):
for k in range(n):
c[i][j] += a[i][k] * b[k][j]
return c

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 54
Exercise
def matrixMul(a, b):
return a@b

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 55
THANK YOU
for YOUR ATTENTION

Huy Tien Nguyen – Tung Le MTH083 - Advanced Programming for Artificial Intelligence Page 56

You might also like