Skip to main content

Quick Look at Numpy

 What is numpy?

  • NumPy is an open source project that enables numerical computing with Python. Numpy is implemented in C, its performance is faster than a Python list.
  • Document


Core Features

  • ndarray: N-dimensional Array Object, the basic data type of NumPy.
  • Vectorization: Operations (or Calculations) performed on the entire array without explicit loops.
  • Broadcasting: Enables operations between arrays of different shapes (or sizes)


Array Operations and Attributes

1. Array Creation

  • np.array(): Creates an array from a Python list.
  • np.zeros(): An array filled with zeros.
  • np.ones(): An array filled with ones.
  • np.arange(): An array of consecutive numbers.
  • np.linspace(): An array with evenly spaced numbers.
  • np.random: Module for generating random arrays.

2. Array Attributes

  • shape: The array's dimensions/axes (rows, columns).
  • dtype: Data type.
  • ndim: Number of dimensions.
  • size: Total number of elements.

3. Array Indexing and Slicing

  • Basic Indexing: arr[0], arr[1, 2]
  • Slicing: arr[1:3], arr[:, 1:]
  • Boolean Indexing: arr[arr > 5]
  • Fancy Indexing: arr[[1, 3, 5]]
    • Allows selection of non-contiguous (non-sequential) elements at once.
    • Enables extraction of elements based on indices that meet specific criteria.
    • Highly useful for data rearrangement and sampling.

4 .Reshaping Arrays

  • reshape(): Changes the shape of an array.
  • flatten(): Flattens the array to one dimension (1D).
  • transpose(): Transposes the array.

5. Broadcasting

Rules that enable operations between arrays of different shapes:

  1. If the number of dimensions differs, a 1 is prepended to the shape of the smaller array.
  2. In any dimension where the size is 1, the array is stretched to match the size of the other array.
  3. An error occurs if the sizes are different and neither is 1.

6. Array Joining and Splitting

  • np.concatenate(): Joins arrays.
  • np.vstack(), np.hstack(): Stacks arrays vertically and horizontally, respectively.
  • np.split(): Splits the array.

Performance Tips

  • Use vectorized operations (instead of Python loops). 
  • Select an appropriate data type
  • Avoid unnecessary array copies
  • Perform conditional operations using np.where().

Labels:

Comments

Post a Comment

Popular posts from this blog

ML Zoomcamp 2025 : Week 1. Intro to Machine Learning

Week 1. Intro to Machine Learning Machine Learning Definition A process of extracting patterns from data An algorithm that learns patterns from data and predicts outcomes. Related concepts Model:  The output of a machine learning algorithm after it has been trained on data. A model encapsulates all the patterns learned during training. Model Training 2 Types of Data - feature: all information about the object - target: What we want to predict about the project Comparison: Rule-Based Systems VS Machine Learning Example: Spam Mail Case 1. Rule-Base System Observing data and create rules to filter spam mail ex) if a mail contains ‘review, promotion, …’ then this mail is a spam -> but spams are keep changing…. it is hard to maintain the code and add new rules… JUST USE MACHINE LEARNING! Case 2. Machine Learning Get Data Define & Calculate features Train and use the model - in this case, use model to classify messages into spam or not spam Process of ML Features: We put t...
Post a Comment
Read more

Datastructure & Algorithm: Tree

 What is Tree? Tree is a data structure which has a hierarchy system.   Concepts & Words Node: each element of the tree Root: The uppermost node of the tree Parent: a node that is above another node Child: a node that is under another node Leaf: Nodes have no children Edge: The line connects nodes Level: The depth from the root (The level of the root is 0) Height: Max level of the tree Features of Tree Hierarchy System: the relationship between subordinates and superiors No cycle: you can't go backward Connections: All nodes are connected from the root Distinct route: There is only one route between two random nodes Implementing the Tree Core Elements to implement 1. Node Class: Each node has the reference to its child node and data 2. Main Functions Insert: Adding a new node Search: Finding a specific value Traversal: Visiting all nodes Delete: Removing a node 3. Ways Of Traversing A Tree Preorder: Root -> Left -> Right Inorder: Left -> Root -> Right Post...
Post a Comment
Read more