The Machine Learning From Scratch repository serves as a valuable companion code resource for the book Machine Learning From Scratch. This repository includes the implementation of ten core machine learning algorithms developed from the ground up using NumPy. Additionally, it presents comparisons with established libraries such as Scikit-learn and PyTorch, enhancing understanding of each algorithm.

Overview of Features

This project demystifies the often opaque processes in machine learning, particularly the functionalities behind fit() and predict(), and provides a comprehensive five-stage framework for algorithm implementation:

1. Intuition: Grasp the conceptual thinking that underpins algorithms in an accessible manner.
2. Formalization: Convert intuitive concepts into mathematical expressions suitable for implementation.
3. Implementation: Translate mathematical formulas into clean, well-documented code using NumPy.
4. Test: Validate the code against real data and benchmark it against popular libraries in the field.
5. Tips: Gain insights into the strengths, weaknesses, and practical applications of each algorithm.

Table of Contents

The repository offers organized chapters covering essential topics and corresponding code implementations, including:

Chapter	Content/Code
Setup	Overview
Fundamentals of ML	Overview
Introduction to Data	Overview
The Math You Actually Need for ML	Overview
Data Preparation	Overview data_loader.py
Linear Regression	01_linear_regression.ipynb
Logistic Regression	02_logistic_regression.ipynb
Regularization	03_regularization.ipynb
K-Nearest Neighbors	04_k_nearest_neighbors.ipynb
Naïve Bayes	05_naive_bayes.ipynb
Decision Tree	06_decision_tree.ipynb
Random Forest	07_random_forest.ipynb
Gradient Boosting	08_gradient_boosting.ipynb
XGBoost	09_xgboost.ipynb
Neural Network	10_neural_network.ipynb
Model Optimization	Overview 11_model_optimization.ipynb
Conclusion	Overview

Core Algorithms

A highlight of this repository is the implementation of ten fundamental ML algorithms:

• Linear Regression
• Logistic Regression
• Regularization
• K-Nearest Neighbors
• Naïve Bayes
• Decision Tree
• Random Forest
• Gradient Boosting
• Extreme Gradient Boosting (XGBoost)
• Neural Network

Practical Applications

Accompanying this code are practical sections guiding how to approach machine learning problem-solving effectively. These sections emphasize the importance of data quality and provide methodologies for model evaluation and optimization. Overall, the repository not only helps in understanding machine learning algorithms but also aids in practical implementation, making it an essential tool for anyone aiming to deepen their knowledge in machine learning.