The machine learning book Amazon sells for about $56 is the same book its authors give away as a PDF. An Introduction to Statistical Learning sits on statlearning.com as a complete free download, in both the R edition and the Python edition. Five of the ten best machine learning books below are free from their own authors in full.
So you do not need help finding good titles. You need to know which ones earn paper, which edition to buy, and who is actually selling the copy you click on, because seven of the listings on this page now have a marketplace seller holding the buy box rather than Amazon. We re-pulled every price, rating count, stock state and buy box seller from Amazon’s catalogue on 30 October 2026, and checked every free PDF on the author’s own site the same week.
Pick a machine learning book by the job you have
If you buy one machine learning book, make it An Introduction to Statistical Learning. It covers the whole classical toolkit in 622 pages, the labs run in Python or R, and you can read the full PDF before you spend a cent. The table gives every pick its job, what the author gives away, and a price band from the 30 October pull, with a direct link in each row.
| Award | Book | Why it wins | Free from the author | Price band | Buy |
|---|---|---|---|---|---|
| Start here | An Introduction to Statistical Learning (Python) | Whole toolkit, no measure theory, Python labs | Full PDF | About $56, R edition about $39 | Check price |
| Best for writing code | Hands-On Machine Learning with Scikit-Learn and PyTorch | 875 pages of projects on the framework people deploy | Notebooks only | About $67 | Check price |
| Desk reference | The Elements of Statistical Learning | The book people cite to settle an argument | Full PDF | About $70 | Check price |
| Maths gap filler | Mathematics for Machine Learning | Linear algebra and calculus tied to real models | Full PDF | About $49 | Check price |
| Graduate breadth | Probabilistic Machine Learning: An Introduction | One probabilistic frame for every model family | Draft PDF, CC BY NC ND | About $86 | Check price |
| Deep learning entry point | Understanding Deep Learning | 68 exercise notebooks, PDF refreshed February 2026 | Full PDF | About $66 | Check price |
| Deep learning in print | Deep Learning: Foundations and Concepts | Bishop’s replacement for his 2006 classic | Browser only | About $51 | Check price |
| Bayesian thinking | Statistical Rethinking | Teaches inference without recipes | Two sample chapters | About $113 | Check price |
| Why any of it works | Computer Age Statistical Inference, Student Edition | The history and reasoning behind the methods | Full PDF, personal use | About $43 | Check price |
| Shortest way in | The Hundred-Page Machine Learning Book | The whole map in a week | Chapter drafts | About $50, hardcover | Check price |
Nobody needs all ten. Most readers need one of the first two rows, plus one more book when a specific gap opens up. That gap is usually the maths, deep learning, or the inference questions that show up the first time you have to defend a model to someone who does not trust it. Prices move daily on Amazon, so treat that column as a band and check the live price before you order.
Check who is selling each copy
We pulled each ASIN from Amazon’s catalogue API for price, list price, page count, publisher, publication date, stock state and the seller holding the buy box. Then we rendered each one on a US listing to read the shipping wording and the rating count, and all ten showed an add to cart button on 30 October.
Where a listing had gone stale or dropped to one copy, the link moved to an edition you can actually buy. The R edition of An Introduction to Statistical Learning links the softcover, Computer Age Statistical Inference links the Student Edition, and The Hundred-Page Machine Learning Book links the hardcover.
The seller check matters more than it did in September. Seven listings named on this page have their buy box held by a marketplace seller rather than Amazon:
- An Introduction to Statistical Learning, R edition
- The Elements of Statistical Learning
- Probabilistic Machine Learning, both volumes
- Understanding Deep Learning, down to a few copies
- Computer Age Statistical Inference, Student Edition
- Deep Learning with Python, Third Edition
That is ordinary for low volume academic hardbacks, and every copy was listed as new. But the seller is the field that decides whether a price holds tomorrow, so each section below says who was selling. Bishop’s Deep Learning went the other way and is back to being sold by Amazon from stock.
We opened the free version claims rather than assuming them. A downloadable PDF, a browser-only reading copy, a draft under a non-commercial licence, chapter drafts and two sample chapters are five different things, and the table above says which one each book has.
Match the book to your level
Pick the row that describes you today, not the one you’re aiming for. The maths column decides whether you finish the book, so be honest with it.
| Where you are | Start with | Maths it assumes | Read next |
|---|---|---|---|
| New to machine learning, want the map | The Hundred-Page Machine Learning Book | Algebra, comfort with notation | An Introduction to Statistical Learning |
| Software engineer, first serious book | An Introduction to Statistical Learning, Python edition | One course in linear regression, no matrix algebra | Hands-On Machine Learning |
| Python developer who learns by running code | Hands-On Machine Learning with Scikit-Learn and PyTorch | Light, explained as it goes | Understanding Deep Learning |
| The maths is what stops you | Mathematics for Machine Learning | School maths | The Elements of Statistical Learning |
| Need a textbook for a course or a reference | The Elements of Statistical Learning | Linear algebra, probability, multivariable calculus | Probabilistic Machine Learning: An Introduction |
| Going straight at deep learning | Understanding Deep Learning or Bishop’s Deep Learning | Calculus, linear algebra, probability | The other one, read free |
| Inference and Bayesian modelling | Statistical Rethinking | Very little, taught through code | Computer Age Statistical Inference |
| Large language models | A different shelf. The LLM engineering books cover transformers in production, retrieval and fine tuning. | ||
| Models in production | Also a different shelf. Start with the MLOps books. | ||
Machine learning textbooks versus practical books
Six of the ten are textbooks in the strict sense: The Elements of Statistical Learning, Probabilistic Machine Learning, Mathematics for Machine Learning, Understanding Deep Learning, Bishop’s Deep Learning and Computer Age Statistical Inference. They carry exercises, they derive what they use, and university courses are built on them. Hands-On Machine Learning and The Hundred-Page Machine Learning Book are practical books, written to get a working model out of you rather than a proof. An Introduction to Statistical Learning sits deliberately between the two camps, a textbook with a lab at the end of every chapter, which is why it keeps winning the first book slot. Statistical Rethinking is a textbook that teaches through code, and it reads more like a course than either camp.
The best machine learning book for Python
Two picks are Python first: the Python edition of An Introduction to Statistical Learning, with its ISLP labs, and Hands-On Machine Learning with Scikit-Learn and PyTorch. Read them in that order. Two more Python books are worth knowing even though neither made the ten, and you’ll find a longer shelf of them in the deep learning and PyTorch books guide.
Sebastian Raschka and his co-authors’ Machine Learning with PyTorch and Scikit-Learn follows the same arc as Géron, scikit-learn for tabular work and then PyTorch for deep learning, across 770 pages. It’s around $34 from Amazon with 529 ratings, the cheapest serious Python option here. The catch is age: it came out in February 2022, before ChatGPT, so transformers get a single chapter and diffusion models are absent.
François Chollet and Matthew Watson’s Deep Learning with Python, Third Edition is the newer of the two, published by Manning in November 2025 as a full rewrite. Chollet created Keras, and the book runs Keras 3 on top of PyTorch, JAX or TensorFlow, then builds a GPT-like language model and a diffusion model from scratch. Print runs around $60 to $70 from a marketplace seller, a price that moved within the day we checked it. The complete text reads free at deeplearningwithpython.io. Read two chapters there before deciding whether it or Hands-On Machine Learning is your one code-first book.
Check the edition before the price
This shelf is full of old editions on sale next to new ones, and the cheaper listing is usually the older text. None of the ten has a newer print edition announced as of October 2026, so the column on the right is what you avoid, not what you wait for, with one exception.
| Book | Current edition | What to avoid or wait for |
|---|---|---|
| An Introduction to Statistical Learning | Python edition, 2023. R users: second edition, 2021 | The 2013 first edition, which has no deep learning, survival analysis or multiple testing chapters |
| The Elements of Statistical Learning | Second edition, 2009 | Nothing newer in print. The free PDF is the corrected twelfth printing from January 2017 |
| Hands-On Machine Learning | Scikit-Learn and PyTorch, late 2025 | The Keras and TensorFlow third edition (2022) is a parallel version, not a mistake. The first two editions are outdated |
| Mathematics for Machine Learning | First edition, 2020 | None. The free PDF carries fixes the print run lacks |
| Probabilistic Machine Learning: An Introduction | 2022, draft PDF dated April 2025 | Murphy’s 2012 Machine Learning: A Probabilistic Perspective, which this replaces |
| Understanding Deep Learning | 2023 print, PDF refreshed February 2026 | None in print |
| Deep Learning: Foundations and Concepts | First edition, late 2023, corrections list July 2025 | Pattern Recognition and Machine Learning (2006), which its own author has now replaced |
| Statistical Rethinking | Second edition, 2020 | A third edition is in progress with no publication date |
| Computer Age Statistical Inference | Student Edition, 2021, with 130 exercises | The 2016 hardback, same text without the exercises |
| The Hundred-Page Machine Learning Book | 2019 | None. The author’s newer book is on language models, not a new edition |
1. An Introduction to Statistical Learning
Gareth James, Daniela Witten, Trevor Hastie and Robert Tibshirani wrote it as the readable sibling of The Elements of Statistical Learning, and Jonathan Taylor joined for the Python edition.
Thirteen chapters take you from what statistical learning is through regression, classification, resampling, regularisation, trees, support vector machines, deep learning, survival analysis, unsupervised learning and multiple testing. Every chapter ends in a lab. The maths is honest but bounded: the authors assume a previous course in linear regression and no matrix algebra, so a working engineer can follow it without stalling.
Two editions are on sale, and the difference is the language the labs run in. The Python edition runs around $56 against a $119.99 list, sold and shipped by Amazon from stock, with 166 ratings. The second R edition is around $38 to $39 with 463 ratings, and a marketplace seller holds its buy box. Same 622 pages, same chapters, so pick by the language you already write.
The Python labs need the ISLP package, and it is still maintained. Version 0.6.0 landed on PyPI on 26 October 2026 and now requires Python 3.11 or newer. On an older distro Python, pip quietly falls back to an older ISLP release, so set up a current Python install and JupyterLab first. For the R edition, the R and RStudio Server setup gives you the equivalent environment.
Buy it if you need to use these methods correctly rather than derive them. Leave it if you already know the difference between ridge and lasso and want the proofs, because this book deliberately withholds them. Both editions are complete free PDFs at statlearning.com, so read a chapter before you spend anything.
2. The Elements of Statistical Learning
Hastie, Tibshirani and Friedman wrote the book that the previous entry politely simplifies. It has 1,361 Amazon ratings, the most of any book here, and it was last revised for print in 2009.
767 pages, and almost none of them are padding. The chapters on model assessment, boosting and the bias variance decomposition are the reason this book refuses to age out. When somebody at work claims a model is overfitting, this is the text that tells you how to prove it.
The free PDF on Hastie’s Stanford page is the corrected twelfth printing from January 2017, eight years of corrections newer than the date on the cover. The print copy runs around $70 and shows in stock, but the buy box moved to a marketplace seller in October, so check who’s shipping it before you order.
The reader this fits wants the derivation, not the recipe. It’s the wrong book if deep learning is your target. The neural network chapter predates the entire modern architecture family, and three books further down cover that ground properly.
3. Hands-On Machine Learning with Scikit-Learn and PyTorch
Géron rewrote the most widely recommended practical machine learning book and changed the framework underneath it. The Keras and TensorFlow edition is still on sale next to it, which makes this the easiest purchase on the page to get wrong.
875 pages, published in late 2025, and the second half is a different book from the third edition. PyTorch replaces Keras, the Hugging Face ecosystem carries the transformer and pretrained model chapters, diffusion models get real coverage, and fine tuning a language model is treated as a normal task rather than an appendix. The notebooks live in the ageron/handson-mlp repository, separate from handson-ml3, which stays on TensorFlow.
The economics changed in October. This edition dropped to around $67 from Amazon, with 85 ratings, while the Keras and TensorFlow third edition sits around $46 with 909 ratings. You are paying roughly $21 for the framework switch, and it is the right $21 if you are learning deep learning today, because PyTorch is what most new work and most published code assumes. O’Reilly keeps both on sale as parallel versions, so buy the older one only if your workplace has already committed to TensorFlow.
Aimed squarely at engineers who learn by running code and want one book spanning classical models and modern neural networks. Where it falls short: theory. Géron shows you what works and moves on, which is the point of the book and also its limit. If what happens after the model trains is your actual job, the MLOps reading list picks up where this one stops.
4. Mathematics for Machine Learning
Most people don’t fail at machine learning. They fail at the linear algebra three pages into a derivation and quietly stop reading, and Deisenroth, Faisal and Ong wrote 398 pages to fix exactly that.
The first half is linear algebra, analytic geometry, matrix decompositions, vector calculus, probability and optimisation. The second half spends that machinery on four problems: linear regression, dimensionality reduction, density estimation and classification with support vector machines. That structure is what separates it from a general maths textbook, because you never learn a technique without seeing the model that needs it.
Cambridge published it in April 2020 and the authors keep the PDF free at mml-book.github.io. The live PDF carries typo fixes that never made it into the printed run, so the free copy is the more correct one. Print runs around $49 from Amazon with 1,051 ratings.
Worth buying the first time you bounce off a proof and want to go back and fix it properly. Not worth it if your linear algebra is already solid, because the payoff chapters are shorter than the setup and you’ll have read better treatments of SVMs elsewhere.
5. Probabilistic Machine Learning: An Introduction
Kevin Murphy’s 864 page rebuild of his 2012 textbook is the widest single volume on this list, and “Introduction” is doing a lot of work in that title.
Everything is framed probabilistically, which is what makes it cohere. Linear models, deep networks, generative models and decision theory arrive as instances of the same idea instead of a tour of unrelated algorithms. That framing is why researchers reach for it and why it’s heavy going as a first book.
MIT Press published it in March 2022. It runs around $86 against a $125 list, with 252 ratings. The listing now shows in stock, but a marketplace seller holds the buy box rather than Amazon. Murphy keeps a draft PDF on the probml site under a CC BY NC ND licence, last dated April 2025. The second volume, Probabilistic Machine Learning: Advanced Topics, adds 1,360 more pages for around $136 and is down to a few copies. Nobody needs both to start.
This one suits graduate students and engineers who want a book that will still answer questions in three years. Two reasons to pass: you are self teaching without a course around you, or you want to finish it, because the density that makes it a good reference makes it a punishing tutorial.
6. Understanding Deep Learning
Simon Prince set out to write the shortest honest deep learning textbook and landed on 544 pages, which in this company counts as restraint.
Chapters are short and each one closes a specific idea: shallow networks, deep networks, loss functions, gradient descent, convolutional networks, transformers, diffusion models, reinforcement learning. The illustrations carry a lot of the explanation. The exercise set is the real differentiator: sixty eight Python notebooks with the code deliberately left out for you to fill in, all runnable in Colab.
The book site hands over the complete PDF, and that copy was refreshed on 9 February 2026, more than two years after the December 2023 print date. Answers to selected questions and an errata list sit next to it. The hardcover is around $66 against a $100 list. A marketplace seller holds the buy box and was down to a few copies on 30 October, so check who’s selling before you order. This is one of the few titles here where the free version is plainly the better artefact.
Good first deep learning book for anyone who has already met gradient descent. It will not help with production engineering, because there is nothing here about serving, quantisation or inference cost. Those questions start in the LLM and AI engineering books and end on a box running NVIDIA drivers and the CUDA toolkit.
7. Deep Learning: Foundations and Concepts
Christopher Bishop wrote Pattern Recognition and Machine Learning in 2006, and a generation of researchers learned from it. This is what he wrote with Hugh Bishop to replace it, and Springer Nature says it was their bestselling title in both 2024 and 2025.
669 pages split into small chapters, with a self contained probability primer so the book doesn’t assume a statistics course you never took. Bishop states the design goal openly: cover the ideas likely to outlast the current architecture cycle rather than chase the last twelve months. Transformers, diffusion models and normalising flows are all there, and nothing is built around a specific library version.
At around $51 against an $89.99 list it is the cheapest serious deep learning hardback here. It is also back to being sold by Amazon from stock after a September stretch with a marketplace seller. The physical object is unusually good too, printed with stitched signatures so it lies flat while you work through the maths. bishopbook.com carries a free to use online version, corrections to the first printing, the full figure set, and exercise solutions for chapters 2 to 10.
Made for readers who want deep learning grounded in probability and intend to keep the book on a shelf. Two reasons to hesitate: worked solutions stop at chapter 10, and the free version reads in a browser without ever becoming a file you own.
8. Statistical Rethinking
Richard McElreath teaches Bayesian inference by taking away the recipes. That is the whole method, and it is why people who already passed a statistics course come out of this book saying they finally understood it.
Causal thinking, priors that you have to justify, multilevel models and posterior predictive checks, with examples in R and Stan. Ports to Python with PyMC and NumPyro are community maintained and easy to find. Nothing else on this list will change how you read a p value in someone else’s report.
It is also the most awkward buy here, and it got more awkward this month. The second edition is from March 2020, and Amazon now lists it around $113 against a $140.99 list price, up from about $93 in September. At least it ships from Amazon’s own stock again. There is no free PDF beyond the first two chapters, and McElreath says on his book page that a third edition is in progress but not coming out right away. His full lecture series is free on YouTube, which softens the price considerably.
Anyone whose job involves defending a number to somebody else should read this. The reason to wait: paying over $100 for a six year old edition while the author writes the next one is a defensible thing to postpone, so watch the price or start with the lectures.
9. Computer Age Statistical Inference
Bradley Efron and Trevor Hastie wrote the book that explains why the field looks the way it does. It answers a question the other nine skip: what happened to statistics when computers arrived.
The book walks from classical inference through the bootstrap, which Efron invented, to the empirical Bayes and large scale testing methods that modern practice runs on, and finally to machine learning itself. Read alongside The Elements of Statistical Learning, it turns a catalogue of techniques into a story with reasons in it.
Cambridge published it in 2016 and the PDF has been free on Hastie’s page since March 2017, with permission, for personal use only. Buy the 2021 Student Edition, not the hardback: it’s the same book in paperback plus 130 class tested exercises, 512 pages, around $43. Stock is thin on both. The Student Edition’s buy box is with a marketplace seller with a handful of copies left, and the 2016 hardback costs more and was down to a single copy. Its 56 ratings are the fewest on this list, which reflects who buys it rather than how good it is.
For the reader who wants context rather than another algorithm. Not the book if you need something actionable this week, because nothing in here tells you what to run tomorrow morning.
10. The Hundred-Page Machine Learning Book
Andriy Burkov compressed the field into a book you can finish on a long weekend, and 1,326 people have rated it, more than eight of the nine titles above it.
160 pages in Amazon’s catalogue, and the compression is the feature. Supervised learning, the main algorithm families, feature engineering, regularisation, evaluation and the practical failure modes all get stated once, clearly, with no worked examples to hide behind. It’s the fastest way to find out which parts of this field you actually care about before committing to an 875 page doorstop.
Burkov sells it on a read first, buy later basis. Chapter drafts download free from the book’s wiki and the Leanpub ebook is pay what you want above a $20 minimum. Check which binding you’re buying, though. The paperback listing currently leads with a used copy from a reseller, while the hardcover runs around $50, sold by Amazon, and is the edition linked here. He has since written a Hundred-Page Language Models Book on the same model.
Best fit: the engineer who has to hold a conversation about a model next week. Its limit is building anything, because 160 pages can’t carry both the map and the terrain, and this one chose the map.
Choose the first book to buy
Four routes cover almost everyone, and none of them starts with buying more than two books.
Coming from software engineering. An Introduction to Statistical Learning first, in whichever language you write, then Hands-On Machine Learning with Scikit-Learn and PyTorch. The first gives you the vocabulary and the reason each method exists, the second turns it into code you’ve run. That pair costs about $105 to $125 at October prices depending on which edition of the first book you take, and it covers everything most jobs ask for. Add Understanding Deep Learning only when a neural network becomes your problem rather than an interest.
Coming from statistics. Skip the introduction and start at The Elements of Statistical Learning, with Computer Age Statistical Inference next to it for the history. Statistical Rethinking comes after that if inference is the part of the job you own.
Going straight at deep learning. Buy one of Understanding Deep Learning or Deep Learning: Foundations and Concepts, never both, and read the other free. Prince gives you 68 exercise notebooks and a PDF that keeps getting updated. Bishop gives you a better physical book, a probability primer, a lower price and, right now, the steadier listing. If you’re unsure, read a chapter of each free before deciding.
Not sure yet. The Hundred-Page Machine Learning Book, one week, free to read before you pay anything. Then come back to this list knowing which of the other nine you need. If what you really want is the big picture of AI rather than a working skill, the books about AI list is the better shelf.
Two books that show up on every other list are deliberately not picks here. Goodfellow, Bengio and Courville’s Deep Learning is free to read in HTML at deeplearningbook.org and remains a fine reference for the mathematics. But it was published in 2016 and predates transformers entirely, so it cannot be a first deep learning book in 2026. Bishop’s Pattern Recognition and Machine Learning is in the same position, and its own author has now written the replacement. Both are worth reading. Neither is worth buying new.
Buy paper only for the books you reread
Five of these ten are complete free PDFs from their own authors, a sixth is a working draft under a non-commercial licence, a seventh reads free in a browser and an eighth hands over chapter drafts. That is unusual for any technical shelf, and it changes what you are actually paying for.
You are not paying for the content. In two cases you are paying for slightly worse content, and in a third for an older printing. The Mathematics for Machine Learning PDF carries fixes that never reached the printed run, and the Understanding Deep Learning PDF was refreshed in February 2026 against a print date of December 2023. The Elements of Statistical Learning free PDF is the corrected twelfth printing, not the 2009 first run, so check the date on the download before assuming your hardback is the current text.
What the money buys is the format. These are books you read with a pen, doubling back three chapters to re check a definition, and page flipping beats scrolling for that. A 767 page reference on a shelf gets opened for years; the same PDF in a downloads folder gets opened twice. So download the free copy, read one chapter properly, and buy paper only for the two or three you keep returning to.
The titles with no free full text still hand you a way to try before you buy:
- Géron: the complete notebook repository
- McElreath: the full lecture series on YouTube
- Burkov: the chapter drafts on the book’s wiki
- Bishop: the free to use online version
Work through one of those for a week. If you keep going back to it, that is the book worth paper.
