Numpy: Array Basics

Creating Arrays

One of the foundational tasks in NumPy is creating arrays. An array is essentially a data structure that stores values of the same type in Python, and it's crucial for machine learning tasks because data in ML is typically represented as arrays. Here's how to create arrays in NumPy:

• From a Python List: You can create a NumPy array directly from a Python list.

  import numpy as np
  arr = np.array([1, 2, 3, 4, 5])
  

• Using Functions: There are several built-in functions in NumPy that allow for array creation, such as zeros(), ones(), and arange().

  zeros_arr = np.zeros(5)  # Creates an array of zeros with length 5
  ones_arr = np.ones(5)    # Creates an array of ones with length 5
  range_arr = np.arange(5) # Creates an array with values [0, 1, 2, 3, 4]
  

Data Types

In NumPy, arrays are grids of values, and they can contain various data types. By default, NumPy tries to guess the data type for the array based on the values you provide, but you can explicitly specify the data type using the dtype parameter.

Common data types include:

• int64: Integer type
• float64: Floating point type
• complex128: Complex number type
• bool: Boolean type (True/False)

For instance, to create an integer array:

arr = np.array([1, 2, 3, 4], dtype='int64')

Slicing & Indexing

Slicing and indexing are vital operations when working with NumPy arrays, especially in the context of machine learning where you often need to select specific data points or split datasets.

• Indexing: Just like Python lists, you can access an array's elements by referring to its index number.

  arr = np.array([10, 20, 30, 40, 50])
  first_element = arr[0]  # Returns 10
  

• Slicing: To select a range of elements from an array, you can use the slice notation, which consists of the start index, end index, and the step value.

  subset = arr[1:4]  # Returns an array with values [20, 30, 40]
  

Reshaping Arrays

The shape of a NumPy array is crucial when you're feeding data into machine learning models. Often, you'll need to reshape data to fit the input or output structure of a model.

The reshape() method allows you to reorganize the data within an array, providing a new shape without changing the data itself.

arr = np.array([1, 2, 3, 4, 5, 6])
reshaped_arr = arr.reshape(2, 3)  # Reshapes the array into a 2x3 matrix

It's crucial to note that the total number of elements must remain the same when reshaping.

Stacking & Splitting

Machine learning often requires merging datasets or splitting them for tasks like training and testing.

• Stacking: You can use vstack() for vertical stacking and hstack() for horizontal stacking.

  arr1 = np.array([1, 2, 3])
  arr2 = np.array([4, 5, 6])
  stacked_arr = np.vstack((arr1, arr2))  # Stacks arr2 below arr1
  

• Splitting: The array_split() function can split an array into multiple smaller arrays.

  arr = np.array([1, 2, 3, 4, 5, 6])
  split_arr = np.array_split(arr, 2)  # Splits the array into two arrays
  

---

By mastering these array basics in NumPy, you'll be well-equipped to handle the data manipulation tasks inherent in the machine learning pipeline. Whether you're preprocessing data, building models, or evaluating results, a solid grasp of NumPy's array functionalities is invaluable.

---

Version 1.0

This is currently an early version of the learning material and it will be updated over time with more detailed information.

A video will be provided with the learning material as well.

Be sure to subscribe to stay up-to-date with the latest updates.

Need help mastering Machine Learning?

Don't just follow along — join me! Get exclusive access to me, your instructor, who can help answer any of your questions. Additionally, get access to a private learning group where you can learn together and support each other on your AI journey.

Subscribe Now