Machine Learning

by Tom M. Mitchell

This textbook provides a single source introduction to the primary approaches to machine learning. It is intended for advanced undergraduate and graduate students, as well as for developers and researchers in the field. No prior background in artificial intelligence or statistics is assumed. Several key algorithms, example data sets, and project-oriented home work assignments discussed in the book are accessible through the World Wide Web.

Several new chapters are available from the author's website.

Genres: Computer Science, Artificial Intelligence, Programming, Textbooks, Technical, Nonfiction, Technology

Hardcover

First published April 30, 1986

Reviews

[Ivan Idris]

This is an introductory book on Machine Learning. There is quite a lot of mathematics and statistics in the book, which I like. A large number of methods and algorithms are introduced:

Neural Networks
Bayesian Learning
Genetic Algorithms
Reinforcement Learning

The material covered is very interesting and clearly explained. I find the presentation, however, a bit lacking. I think it has to do with the chosen fonts and lack of highlighting of important terms. Maybe it would have been better to have shorter paragraphs too.

If you are looking for an introductory book on machine learning right now, I would not recommend this book, because in recent years better books have been written on the subject. These are better obviously, because they include coverage of more modern techniques. I give this book 3 out of 5 stars.

[Anthony Singhavong]

Great intro to ML! For someone who doesn't have a formal Comp Sci background, this took a lot out of me. I found it helpful to stop after every chapter and listen to a more recent lecture to tie loose ends. Highly recommend reading this book in conjunction with professor Ng's ML intro course.

[David]

This is a very compact, densely written volume. It covers all the basics of machine learning: perceptrons, support vector machines, neural networks, decision trees, Bayesian learning, etc. Algorithms are explained, but from a very high level, so this isn't a good reference if you're looking for tutorials or implementation details. However, it's quite handy to have on your shelf for a quick reference.

[Samars]

This book is absolutely amazing. I love so much is my favorite book.

[Lukas Simons]

🙄

[Steve]

Great theoretically grounded intro to many ML topics.

[Jacob Williams]

This is part of the required reading for the machine learning class in Georgia Tech's online master's program. I think it's weird that they use a book from the 90s, and that's especially annoying because it's expensive and relatively hard to find. It is a really well-written textbook, though.

[Daniel Smith]

Really loved this book!

This was my introductory book into the how and why machine learning works. I still come back to this book from time to time to serve as a reference point!

In my opinion Tom Mitchell serves up some good motivating examples for the algorithms and simply and clearly explains how they work.

[Conor Livingston]

I learned a lot from this book. The author assumes very little prior knowledge about math and statistics. For that reason, he takes care to explain equations thoroughly from a rigorous and intuitive perspective.

The book is old, and you'll see many references from 1980s and 1990s. However, the content isn't about any specific technology it's about the foundational ideas in the field of machine learning. For that reason, the content is still relevant in my opinion.

I would recommend this book to anyone machine learning beginner who wants to dive deeper into the field.

[HDubbs]

This books is old, but I thought it was a great introduction to Machine Learning. I‘d never heard of Random Forrests before reading this book and this was very helpful, BUT of you already a lot about ML, consider something else. This is an intuitive introduction for people who have basic math skills or first year grad students (maybe even the motivated bachelor student).

[Abhijit Gupta]

Read it back in 2018. I think it was my first book on ML. if you're well versed in probability, stats & linear algebra and want to get an outline of traditional ML, then this should serve as a nice weekend course on ML. :D

[Cora Meador]

Yes, everything in the book is technically sound, the material holds up even after 30 years, and it is used in all of the machine learning courses. The problem is that the way the book is written, it's hard to parse information out of it, which essentially means that you have to read and follow along with entire paragraphs of boilerplate writing in passive voice with tiny nuggets of information that could have been more easily explained in a table or list. This makes it both hard to learn from, and hard to use as a reference.

[Andrew Miller]

This nook is good for beginners in machine learning. Covers all of the basics pretty well. If you are interested in more modern sources, check https://voicesfromtheblogs.com/ai-ann.... Here you can learn about modern AI annotation tools and other trends in the data processing. Working with raw data at the beginning of building software with machine learning features is really important, so the more you know the better.

[Lurker]

Probably the first book you want in academic setting when studying machine learning. it's simple yet effective, and contains less mathematical mind-twisters and more concepts of machine learning algorithms.

[Jethro Kuan]

A little dated, but had a nice way of introducing machine learning, classifying learning algorithms by their inductive biases. Would recommend other more modern books, such as the one by Kevin Murphy.

[Nick Davila]

I read about half maybe more than half of this book for a class. I skimmed the rest. Crazy how this book is responsible for a lot of the way the world is now. Just some math controlling the world!! Pretty good book though very dense! Good for theory not really much for applications.

[Niraj Upadhayaya]

A very nice introduction to Machine learning. It gives solid foundations. However very few examples are there.

[Sushma Anand]

Awesome book and so simply explained .

[Robert]

My introduction to the field. It's pretty awesome.

[Mdhafeez]

nice

[JR Smith]

The coverage is fine, but this is one of those textbooks that values rigor to the detriment of actually explaining the material well.

[Benjamin Kozel]

A fine high level look at a wide breadth of Machine Learning concepts and algorithms

[Andrew]

Read portions for CS7641 - Machine Learning for Georgia Institute of Technology. De facto standard book for machine learning

Merged review:

Read portions for CS7641 - Machine Learning for Georgia Institute of Technology. De facto standard book for machine learning

[Eric Wallace]

Textbooks like this might not make for "fun" reading, but sometimes they're quite necessary. I was reading Tom Mitchell's classic machine learning survey book along with a machine learning survey class, as you might guess... and found it quite helpful. While the professor for my class made an effort to explain concepts and algorithms and such, I rarely began to understand the lectures until I read the textbook.

One key point for reading this book: each chapter and section builds upon past content, so reading from the beginning is actually a good idea (despite what the intro says about jumping into what chapter you need). Mitchell is always careful to explain terminology and such so you don't necessarily need to look it up elsewhere, so if a term pops up that you're not familiar with, chances are good you missed the earlier explanation of it!

[Brian Powell]

An easy, engaging text with a good selection of introductory topics from the field of machine learning. Mitchell covers decision trees, neural nets, Bayesian methods, rules and concept learning, and reinforcement learning, among others. Each is treated at just the right level: enough detail to chew on the concepts, but not a slog into the marginal particulars of this or that technique. Recommended as a good starter kit for those interested in machine learning: from here, you can launch off in a variety of directions, or supplement with other resources as needed. Only caveat is that it's a tad dated (and no coverage of unsupervised learning), though this is a minor nit for a lightweight introduction to the theory and concepts behind the various learning approaches.

[Chris]

Machine Learning by Tom Mitchell was a good read that was surprisingly light on the math. It covered several different machine learning algorithms including: Concept Learning, Decision Tree, Neural Networks, Bayesian, Genetic Algorithms, Analytical Learning and Reinforcement Learning. It also mentions how to evaluate algorithms providing a training set limit equation and discussed how to evaluate hypothesis using confidence intervals. I enjoyed the structure and re-occurrence of specific concepts that I wasn't familiar with like inductive bias and hypothesis search space. The beginning, which used Concept Learning was very useful as a proxy as the later algorithms get more complex.

[Liuyang Li]

The book does a good job summarizing various research areas in machine learning. It, however, lacks enough formal treatments. It also does not provide enough details for each topics so that the read can fully understand the algorithms to apply them. I would recommend reading this book as an overview of machine learning and read books on each individual topic afterwards.

[Terran M]

This book is a classic, but I personally don't care for it - it strongly embodies the "ML view", which sees the world from the point of view of optimizing a cost function for future examples, without deriving insights from models. In other words, it is all about prediction and not about inference, which is not the point of view that I personally find interesting and valuable.

[Todd Johnson]

Very clear prose. Covers an interesting sample of both probabilistic and non-probabilistic methods. Starting to feel a bit dated, as it does not cover important methods developed over the last decade, such as support vector machines. Nonetheless, the topics covered are covered very well.

[Frank Palardy]

This is a little harder than the Russell AI book, but it doesn't have the problem of not being complete. It doesn't try to give simple examples, but the math is complete, or complete enough for me. I didn't finish it all. It's more like a reference.