NumPy Intro (Google Machine Learning Course)
September 14, 2020 by Jane

Importing the NumPy module

import numpy as np 

 

Populating Arrays

array = np.array([1.2, 2.4, 3.5, 4.7, 6.1, 7.2, 8.3, 9.5]) // create 1D array

matrix = np.array([[6, 5], [11, 7], [4, 8]]) //create 2D array

seq =  np.arange(5, 12) // create an array consisting of integers 5, 6, 7, 8, 9, 10, 11 - the range is [a, b)

randints = np.random.randint(low=50, high=101, size=(6)) // create an array of 6 with random integers  from 50 - 100 

randfloats = np.random.random([6]) //default is to create random floats between 0 and 1

array_zeros = np.zeros([6]) // populate an array with 6 zeros

array_ones = np.ones([6]) // populate an array with 6 ones

 

Broadcasting:

A feature of NumPy that when two matrices are not the same size for the intended operation, it will expand the smaller operand to dimensions that can be operated on for linear algebra.

E.g. randfloats = np.random.random([6]) + 2.0

  • broadcasting will expand 2.0 to [2.0, 2.0, 2.0, 2.0, 2.0, 2.0]