The Hundred-Page Machine Learning Book (The Hundred-Page Books) Paperback – January 13, 2019

When I saw that the book was laid out in LaTeX, I knew I was in trouble.The order of introduction of ideas is haphazard and over-detailed without ever providing an over-arching context by which to interpret the minutiae or determine if they are actually important to the reader's goal in reading. In short, it reads like it was written by a statistician or computer scientist who simply does not know how to communicate new information to a naive audience. People like Burkov are the reason that so many suffer from the misconception that they are bad at math, when it's actually the case that math people suffer from the misconception that they know how to explain it.Here's how a book like this should be laid out:Ch. 1: Applications of ML.—This provides a series of example problems that the reader can conceptualize, drawn from various fields, but none so detailed that readers from outside those fields cannot follow them. For each, it lays out the difficulty in answering the question at hand, then walks through what would need to be done to do so, and finally gives a quick thumbnail sketch of what ML algorithm could do it, and how.Ch. 2: Fundamentals of MLThis outlines the basics of any of these approaches—what their data look like and what the outputs are. Equations are included, but instead of being simply typed out in LaTeX and then discussed in prose, all the terms are listed beneath the equation with a quick description and example of each. Even better, they could be presented as figures and labeled with same. This chapter should also cover the "Anatomy of a Learning Algorithm," something not covered in this text until Ch. 4, after the minutiae of various algorithms and models have been discussed without any scaffolding.After that, the book should have detailed chapters on each of the major families of algorithms.The end of the book should focus on practical aspects of using the algorithms.It's maddening to see such a poor attempt at pedagogy. It really isn't that hard to explain technical concepts to beginners; you just have to start top-down to contextualize, frame, and scaffold the topic, then you can drill down into specifics. You don't start bottom up. No one learns physics by starting with quantum theory, for crying out loud.The positive reviews of this book are likely from people who were already in closely-related fields. If you're from another field—even one that is quantitative, and even one that uses some of the models employed by ML algorithms—you're going to spend a lot of time on Wikipedia while you "read" this book, trying to figure out if you even care about the current passage or not.It's bad.

Not suitable and enough for beginners.

My professor recommended me to buy this book. It is very useful to my study. No different words and easy to understand. Even the texture of book page hahaha