Machine Learning by Hand
Core Algorithms Worked Through With Small Numbers
- Pages
- 82
- Chapters
- 15
- Level
- Intermediate
- Format
- E-book
About this book
Most machine-learning books either hide the mechanics behind a library call or bury them in notation. This book does something different: it works through each core algorithm by hand, with datasets small enough to fit on a page, so you can see exactly what happens at every step.
You will fit a regression line and calculate its error, run gradient descent for several iterations, compute a logistic-regression probability, choose the best split for a decision tree using Gini impurity, and watch k-means move its centres until they settle. Along the way you will learn the ideas that matter in every real project: features and labels, loss functions, validation, overfitting, regularisation and evaluation metrics.
The final chapter ties everything together in a complete project walkthrough, from question to deployed model. Only school-level algebra is required; every formula is explained in words first.
What you will learn
- How data is turned into features and labels
- Linear regression, loss and the meaning of 'best fit'
- Gradient descent, computed step by step
- Logistic regression and decision trees for classification
- Overfitting, cross-validation and regularisation
- k-means clustering and a complete end-to-end project
- k-nearest neighbours, naive Bayes and gradient boosting, each worked by hand
- A full forward and backward pass through a neural network, with real numbers
- ROC curves, AUC, PCA and feature engineering, plus practice problems with solutions
Table of contents
- IntroductionIntroduction
- Chapter 1Data, Features and Labels
- Chapter 2Linear Regression and the Idea of Loss
- Chapter 3Gradient Descent by Hand
- Chapter 4Classification With Logistic Regression
- Chapter 5k-Nearest Neighbours
- Chapter 6Naive Bayes
- Chapter 7Decision Trees and Forests
- Chapter 8Gradient Boosting by Hand
- Chapter 9A Neural Network by Hand
- Chapter 10Overfitting, Validation and Regularisation
- Chapter 11Measuring Models Properly
- Chapter 12k-Means Clustering
- Chapter 13Principal Component Analysis, Gently
- Chapter 14Feature Engineering
- Chapter 15A Complete Project, Start to Finish
- AppendixPractice Problems
- AppendixA Maths Refresher
- AppendixFrom Hand to Code
Who this book is for
Students, analysts and developers who want to really understand how classic machine-learning algorithms work before (or while) using libraries. School-level algebra is enough.
Details
| Format | PDF + web edition |
|---|---|
| Length | 82 pages (approx.) |
| Chapters | 15 |
| Level | Intermediate |
| Language | English |
| Topic | Fundamentals |
| Author | Latent Library Editorial |
Keep reading