Python Machine Learning, 1st Edition

In my opinion this is a great book to get you up and running with machine learning. It manages to not only cover the basics but also talks about some of the more advanced topics. There are a couple of things that I really liked about this book. 1. You learn a lot of things that you can't find online and that are APPLICABLE to the real world. Even if you just want to get into machine learning and use it but don't necessarily want to become a data scientist this is a great buy. Machine Learning can be really useful when put into good use. I for example, after reading the book, was able to quickly write up a python program to predict what time I would wake up based on what time I slept, what day it was etc. As well as having tons of fun playing with data from http://archive.ics.uci.edu/ml/ . 2. Although this book is focusing on python the math that you need to implement the algorithms are all there. What's great about that is that I was able to "Translate" most of the examples from the book to C++ code without much hustle. Not only that but the math behind these algorithms made a lot more sense after reading this book. So even if you don't necessarily want to use python but want to gain intuition over how these algorithms work this book will also come in handy. 3. This book isn't just about Machine Learning algorithms. It actually talks quite a bit about preparing and getting good data in general. Which is crucial for every data scientist since almost 80% of your job is getting good data. And another 20% finding a good model and training it. Overall I would say that this book helped me and that I learnt a bunch of new things. If the above didn't convince you (Along with other reviews) here are some small details that made the reading of this book a joyful experience. - I felt that reading the book was actually really fun and motivating since at every chapter there were several examples of applying the theory taught. Which motivated me to move on and read more. - Although this may not seem as important. I have to say that the font of the book as well as the tone of the writing made the reading of the book really comfortable and joyful. I didn't feel that I was getting tired and was easy for me to pick it up where I left off. I have to say though that there where some typos here and there (I thing I found 2-3 in total as well as 2 pictures where swapped) but they were easy to spot so it wasn't that big of a problem. Reasons why you shouldn't buy this book: Unless you are a Machine Learning expert and you look into the deeper insights and more advanced stuff in Machine Learning you shouldn't be looking into buying this book since most of the stuff taught is already known to you. (Although I doubt that you would be looking through this reviews thinking whether to buy it or not in this case). I have also included some pictures. Great Book. Highly Recommend it!

This is a fantastic book, even for a relative beginner to machine learning such as myself. The first thing that comes to mind after reading this book is that it was the perfect blend (for me at least) of theory and practice, as well as breadth and depth. Let’s face it, we know that machine learning isn’t an easy subject. You need theory…but you also need practice in the form of some serious coding before you really start understanding it. And this is one area where Sebastian’s book shines: it contains a plethora of really good code examples that are illuminating and well explained, and which cover a very wide range of different machine learning algorithms. And, speaking of code, as another reviewer has pointed out, another huge plus is that, in many places, Sebastian shows you how to gauge the performance of your code and make it more efficient. For me, the best measure of any book such as this is how many “ah ha!” moments I had while reading it. And I had more than a few while reading Sebastian’s book. One such “ah ha!” moment came while reading chapter 12 (and this also illustrates that nice blend of theory and practice I already mentioned above). In this particular chapter, he discusses training artificial neural networks for image recognition. At the heart of this approach is back propagation, which is pretty much THE bread and butter behind multilayered neural networks. He presents a detailed discussion of back propagation in two separate pieces: one that is intuitive and “top down”; the other a more mathematical, “bottoms up” approach that goes through the algorithm step by step, showing how the gradients are computed and the weights updated. His treatment of back propagation was one of the better explanations I’ve seen and really cleared things up for me. One last thing I must mention: at the time of release, this was the first machine learning book for Python (to my knowledge) that has an entire chapter devoted to Theano, which he uses to parallelize neural network training. For those who don’t know, Theano is a particularly nice (not to mention very powerful) Python library for doing machine learning, most especially if you can utilize the power of GPU computing. In addition, that particular chapter (13) also introduces the brand new Python library named Keras, which is built on top of Theano and is a really nice library for the rapid building and prototyping of neural networks (in the spirit of Torch). Being a brand new library, his treatment of Keras was necessarily brief, but it was a great starting point. In conclusion, I am very confident that if you do pick up this book, you won’t be at all disappointed. And be sure and grab the accompanying code for the book on his GitHub repository (just look for “python-machine-learning-book” on github.com/rasbt.) His code is top notch and I’ve yet to encounter any problems with it.

Update 11/2/2015: Finally got the replacement from Amazon. This one is in perfect shape, i.e., no alignment issues at all. Woohoo! Just as promised. For other details, please see my 10/29/15 review below. Update 10/29/2015: After having a conversation with the author (see comments), I'm updating my review. Here are some updated remarks on my original comments: I've been pushed a new copy from Amazon and promised properly aligned pages. I took 2 stars off for the misaligned pages (I'm emailing a picture of these to the author), but I'm giving these stars back because - as per our conversation - my copies were outliers and hopefully the new copy I receive will be properly printed. The three copies the author has are printed properly and mine shouldn't have been different. I'm also adding a star back because the author explained how the publisher/graphics team didn't do their job right. In conclusion, I'm taking off just one star because the whole 'package' I received isn't how I expected to be, but this isn't a poor reflection on the author. Actually, it's quite the contrary, and I stand by my statement that his content, ideas, thoughts, code and plots do their job well. The breadth of information is immense, and the depth on each topic is enough that I would consider this a great scikit-learn reference. Original review 10/26/2015 (deprecated): I didn't want to give this a bad review, trust me. I believe the content is great, the author has done a good job, and he's clearly knowledgeable on the subject matter. The content seems strong, covers a wide-variety of topics, and is a great reference on the Python scikit-learn package. I'm not sure one would be able to learn machine learning from this book alone, primarily because it's actually a reference on scikit-learn, but also because the algorithms are not implemented from scratch and instead covered in concept (and implemented in sklearn with examples here and there). As a reference to sklearn, this book does its job well. My issue is with the publisher and the editors. EVERY page is printed misaligned, i.e., not printed parallel to the paper. More alarmingly, opposite pages are misaligned in OPPOSITE directions, giving the perception that the pages are misaligned more than they really are! It's horribly distracting while reading the content. I thought it was a one-off thing, but I have seen this in two copies so far. I took off 2 stars for this. The other issue is with the formatting of the book. Have a look at the contents, for example; there is no line-spacing before or after chapter titles. Going through the contents section feels like I'm reading strange prose with italics, bold and underlined text thrown in here and there. I have to pretty much actively parse the entire contents to find what I'm looking for, or give up in the contents and jump back to the index. It's an annoying, utterly unsatisfying experience and pushes me away from the book. I took a star off here. The final thing wrong with this book are the illustrations. The charts/plots themselves are okay; they seem to be implemented in matplotlib, our go-to Python plotting library. They do their job, but definitely lack a certain je ne sais quoi. I guess they could've been prettier (think Bokeh or ggplot plotting libraries), but like I said I'm neutral on this. My problem is with the conceptual illustrations. That is, when the author is explaining machine learning or data science concepts, he seems to rely on what looks like MS PowerPoint 2003 or older to build out his diagrams. These plots are so 2003 that I have a hard time gleaning the concepts they're trying to push across. I'm sure the editors could've easily spent some more time recreating these charts in a newer version of PowerPoint (maybe with 2010+ design principles), or even better, in an actual graphic-design platform. Imagine browsing webpages on a Palm Pilot after using an iPhone/Android device for the last decade or so. You're not going to like it, it's going to slow you down, and maybe it'll even impede learning; this is what the illustrations are like. Ugh - I really wish the author, publisher or editor hadn't used the default template on PowerPoint 2003. Puts a bad taste in my mouth every time I think about it. Minus one star for this.

I'm a senior undergraduate student in electrical and computer engineering, and decided to make use of machine learning for my senior design project. Having had some experience in python (but not much with matplotlib or scipy), I decided on this book because of all the other good reviews. I have since picked up a few other books related to machine learning, but none can even compare to this. It's stellar! In three weeks I have managed to give myself a comprehensive crash course in classification algorithms using Python which is enough to give me a rolling start on my design project. I am about half way through the book, and apart from very few minor errors (to be expected in a first edition book), I cannot find any faults in it. It's a great resource for someone who wants to learn about machine learning but doesn't know where to start. I'm going to keep an eye out for further books by Mr. Raschka, because his ability to clearly and concisely explain things is superb. Additionally, I enjoy the fact that the book attempts to give a solid foundation on the mathematics behind various machine learning algorithms, since that is enjoyable for someone like me, who always likes to understand what is happening beneath the surface. Update: Having finished the book now, I can definitely reaffirm my original position. This is one of the best technical books I have ever read. The last few chapters especially, image recognition with MLP networks and parallelizing networks with Theano and Keras are extremely interesting. I have taken these ideas and applied them in several of my own projects now. Also, as I'm planning on going to graduate school in the very near future, I'm thinking that machine learning and ANNs will likely be at the top of my list of areas to specialize in. The research that is going on in this field is huge, and this book manages to touch at the very base of neural networks, but enough to get your feet wet and show you where to go from there.

I would not recommend the Kindle edition because of the formatting of the book. All the equations, inline and standalone, are too small to read. The graphs are too small to read. You have to double click on each one to zoom in. Then you click to go back to page. This horrible formatting makes reading the book tiring and frustrating. The book gets a 1 star for the page layout making it difficult. I don't understand how the publisher let this go. I will check samples carefully in the future to be sure not to get stuck with one of these duds again. I would advise anyone to try the epub format straight from the publisher. I can't say that is better, but can't imagine it could be worse. The content was okay, but only left me looking for another book on the subject. The examples are excellent working code case studies. The balance of theory vs practice in the book wasn't right for me. I want to use Machine Learning. I do not intend to advance the field. While the math is interesting, it doesn't help a user decide what to do nor how to apply a method to a problem. Too much of the advice in the field seems to reduce to "try it, vary the parameters, maybe it will do something for your problem" or "if that one doesn't work, here are others to try". I realize that this may be state of the art, that people don't really know _why_ the various machine models work or don't work. But to move machine learning from a science experiment to a methodology with predictable outcomes, we need more books focused on practical advice.

I'm a data scientist and I've read and actually worked the examples in at least 5 data science books looking for good reference/introduction material for my interns. This book is by far the best book I have come across. It's put to good use here in the office and the interns love it - some actually going so far as to purchase their own copies. To fully benefit from all it has to offer you'll need a solid background in mathematics(calculus and linear algebra) and python (or some other object-oriented programming language). This is not a beginner's book, and I'm confused by people who are surprised at that - it says as much in the preface. With the requisite background comfortably under your belt, you'll find that this book not only covers the machine learning algorithms popular in data science today, but also delves into other related and very necessary topics such as preprocessing, data compression, and parameter tuning. Too many data science books today give an overview of the algorithms themselves (albeit sometimes a very deep overview), but offer no (or very little) instruction on how to use them in a practical sense, as part of a workflow of some kind. This book does both beautifully. Any book of this kind will contain errata. The only bad thing is that the publisher does a bad job of documenting them. Luckily the author keeps an updated and complete list on github. I suggest you go through the errata and make corrections in the book before you start reading it. There aren't too many, but it will keep you from getting stuck when your code doesn't run like it should. [...] Anyway...buy this book. I highly recommend it to anyone with the requisite background trying to get a better understanding of data science and machine learning.

I've read through this book multiple times now. The first time I skimmed it as I needed it and used it more as a reference. It was very helpful and got me going in the right direction. Now that I've developed a deeper understanding of machine learning, especially in python, I'm in the process of reading through the entire book from the beginning. I must say that this is an excellent book and it's a great way to deepen your understanding of machine learning. Both in how algorithms work (as the book starts off with a very basic example of the perceptron model, makes it accessible, and builds from it), and how to apply effective and practical machine learning. There really is no one-stop-shop, and take the advice from the author that there is no free lunch. Nonetheless, when it comes to compiling a bunch of useful, accessible, and practical machine learning knowledge, the author did a great job at it. If you take your time to work through this book and ensure your understanding before moving on, then it will prepare you for making things like research papers and other more advanced machine learning topics more accessible . If you think this book is too hard then you aren't prepared to step into the world of machine learning.

One of the very good books on Machine Learning. What I liked: -good review of main algorithms in Machine Learning -handy code in Python -Book designed with the practitioner in mind -Will have you covered for a good level -addresses some of the popular libraries (Scikit learn for ML, Theanos and Keras for Deep Learning) What I found less good: -I already know most of the algorithms for having written them during different courses so I found it a bit long to have to go through them all over again (but others, more detailed oriented may find the actual implementation interesting) -Sometimes, the book attempts to cram in too much (the portion about Flask web design is a bit pushing too much in another direction) -the part on Deep Learning and signal processing is a bit too light for my taste even if in some ways above the crop (especially for the Theanos and Keras parts that provide a good introduction to these libraries). All in all, it is a good book on practical Machine Learning that delivers what it's title says.

I have been an ML practitioner for years. The majority of my time has been spent on deducting formulas and work with stats models. I like this book as it provides some great tips for ML production in Python. Before reading the book, I did not know some of the utility functions, such as stratified k-fold, are already there in sklearn. Because I do not worry about the theory and the implementation, I quickly flew through the book in days and learned some interesting points. I would recommend this book to the software engineers/developers who want to start a career in data science. It may not be a good one for research community as at many points the discussion could be superficial. However, this makes sense as the depth is not the focus of the book:) One improvement I expect from the next version(if possible) is the color -- b/w makes the figures extremely hard to follow.

No me interesa adquirir el producto

Great intro to machine learning algorithms. Since the author focus mainly on algorithms (using Python's scientific libraries), the explanations may be non-mathematicians friendly.

E' stato il mio primo approccio al Machine Learning, avendo una base di matematica e statistica a livello universitario e di programmazione in Python per applicazioni scientifiche (Numpy, Pandas, Scipy, Matplotlib). L'ho trovato molto chiaro e molto bello. Credo sia utile anche per coloro che vogliano approfittare per imparare a lavorare in Python. Gli ultimi 2 capitoli riguardano il deep learning e sembra esser un po l'introduzione di un altro libro da studiare...

Great Introduction to Machine Learning with Scikit-Learn! Very well written, lots of examples. Very suitable for machine learning beginners with python experience!