What do you think?
Rate this book
Building Machine Learning Systems with Python
As the Big Data explosion continues at an almost incomprehensible rate, being able to understand and process it becomes even more challenging. With Building Machine Learning Systems with Python, you'll learn everything you need to tackle the modern data deluge - by harnessing the unique capabilities of Python and its extensive range of numerical and scientific libraries, you will be able to create complex algorithms that can 'learn' from data, allowing you to uncover patterns, make predictions, and gain a more in-depth understanding of your data. Featuring a wealth of real-world examples, this book provides gives you with an accessible route into Python machine learning. Learn the Iris dataset, find out how to build complex classifiers, and get to grips with clustering through practical examples that deliver complex ideas with clarity. Dig deeper into machine learning, and discover guidance on classification and regression, with practical machine learning projects outlining effective strategies for sentiment analysis and basket analysis. The book also takes you through the latest in computer vision, demonstrating how image processing can be used for pattern recognition, as well as showing you how to get a clearer picture of your data and trends by using dimensionality reduction. Keep up to speed with one of the most exciting trends to emerge from the world of data science and dig deeper into your data with Python with this unique data science tutorial.
350 pages, Paperback
First published January 1, 2013
43 people are currently reading
457 people want to read
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 13 of 13 reviews
December 10, 2014
Quite an impressive book. I’m not sure what has amazed me most: the examples, so real and complex but still, easy to follow, or the easiness through which the authors introduce the reader to many different machine learning techniques without throwing at him scaring math formulas. Certainly both things. And not only those.
As clearly stated in the preface, this book it’s not about scikit-learn. It does not expect the reader to master it. Similarly, it does not aim to teach him the library from install to Z. The book is about Python and machine learning. Many examples are, indeed, pure Python.
Still, scikit-learn is used. Sometimes alone, often in combination with other widely known libraries, such as Numpy, Scipy, nltk and matplotlib. While scikit-learn and nltk are not taken from granted, whoever is holding the book on his hands is expected to have been exposed already to both the numerical libraries and to matplotlib. Methods such as norm() and rvs() are not explained and if you are not familiar with them, get ready to open both the reference and a terminal with the Python interpret.
Not only do the authors rely on those must know libraries. They also show the readers some less known gem, such as gensim (a real beauty that doesn’t ship with scikit-learn).
As I was saying at the beginning, the examples are very interesting. They are real world challenges, clearly explained step by step with many accompanying charts and schemas. The source code that comes with the book has some snippet that you really wanna copy somewhere safe.
I have particularly enjoyed the idea behind the book itself: introduce the reader to machine learning without exposing him to the mathematical details behind it. The concepts are very well explained and the few formulas you will see won’t get you screaming terrified. I pretty much agree with this approach: we don’t need to know how an engine works to drive a car.
I do absolutely recommend this little masterpiece to any Python programmer who wanna get started with machine learning. This book really get you curious about the subject. Isn’t it what all great books do?
As usual, you can find more reviews on my personal blog: http://books.lostinmalloc.com. Feel free to pass by and share your thoughts!
As clearly stated in the preface, this book it’s not about scikit-learn. It does not expect the reader to master it. Similarly, it does not aim to teach him the library from install to Z. The book is about Python and machine learning. Many examples are, indeed, pure Python.
Still, scikit-learn is used. Sometimes alone, often in combination with other widely known libraries, such as Numpy, Scipy, nltk and matplotlib. While scikit-learn and nltk are not taken from granted, whoever is holding the book on his hands is expected to have been exposed already to both the numerical libraries and to matplotlib. Methods such as norm() and rvs() are not explained and if you are not familiar with them, get ready to open both the reference and a terminal with the Python interpret.
Not only do the authors rely on those must know libraries. They also show the readers some less known gem, such as gensim (a real beauty that doesn’t ship with scikit-learn).
As I was saying at the beginning, the examples are very interesting. They are real world challenges, clearly explained step by step with many accompanying charts and schemas. The source code that comes with the book has some snippet that you really wanna copy somewhere safe.
I have particularly enjoyed the idea behind the book itself: introduce the reader to machine learning without exposing him to the mathematical details behind it. The concepts are very well explained and the few formulas you will see won’t get you screaming terrified. I pretty much agree with this approach: we don’t need to know how an engine works to drive a car.
I do absolutely recommend this little masterpiece to any Python programmer who wanna get started with machine learning. This book really get you curious about the subject. Isn’t it what all great books do?
As usual, you can find more reviews on my personal blog: http://books.lostinmalloc.com. Feel free to pass by and share your thoughts!
January 22, 2014
This is one of those books where the concept is good but the execution leaves much to be desired. I like the idea of a book of case studies: datasets, machine learning techniques, Python - all mixed together to give a practical, hands-on treatment. But there are problems with the organization of the material, there are inaccuracies in definitions; unevenness in the depth of exposition; and problems with some of the code examples.
Nice idea, but needed much more care.
Nice idea, but needed much more care.
September 17, 2019
Not bad, but very outdated for Sept 2019.
June 21, 2020
Un buen libro de incidencia, los algoritmos y están bien implementados y es muy funcional
September 23, 2013
This is a twelve chapters tutorial on an increasingly important subject, Machine Learning, through the use of one the most widely used and general purpose computational languages, Python.
http://www.packtpub.com/building-mach...
The authors assume we hve already installed at least a Python 2.7 version in our machine. Then, in chapter 1 (chapter, hereafter Ch) they promptly explain the main lines of the book, what we could learn from it and how/when/where to apply this knowledge. Then they focus in the Python tools/libraries needed:, Numpy, Scipy, Matplotlib and Scikit-learn, and how to use them to get the most of our data. They reveal the power of those libraries: a particularly nice example is given at the very beginning when they compare runtime behaviours. They end the first chapter with a full fledged example: goal, methods, needed tools and vocabulary with words as overfitting, test and training, in the context of machine learning.
In Ch 2 they move into real life examples and the use of ‘external’ (though, classical) datasets and the notion of ‘classification’ first through the visualization of the dataset and then through coding for the extraction of features, construction of models (able to be used with different, new datasets), their evaluation, and more complex classifiers. Ch 3 is about Clustering (Flat and Hierarchical) and the finding of Related Posts in, say,a blog website, through the search and pre processing of similarities between common words using the notion of words stem and how to deal withy the related noise. In Ch 4 the authors start the discussion by taking this book as an example on how to classify the book (machine learning or python?) expanding from clustering to the notion of topics. Cleverly, they introduce a couple of concepts and the acronym LDA both as an exercise as well as important definitions: we learn on every page! In Ch 5 they are back into classification and teach us how to detect poor answers example, through the preselection and processing of attributes and defining what is a good answer and how to deal with variances and bias, accuracy and precision. Ch 6 is about …. Sentiments and (example) Twitter opinion mining, critical for commerce and politics through the use of the Naïve Bayes classifier (and Theorem) one of “the most elegant machine learning algorithms”. In Ch they give us some Recommendations on Regression plus the notion of hyperparameters. Ch 8 provides even better ideas and methods for Regression and Recommendations and what happens when there are not numerical ratings, introducing the Basket Analysis method (example, Amazon “customers who bought this also bought…”) applicable when we want to to do a recommendation example, based on predictors and association rules. Ch 9 on Music Genre Classification, not by, say, titles, but actually inspecting the data, example, the spectrograms and decomposition on wave components. Hardly could youu imagine it more technical, inspirating and useful! This allows to go into more specifics as the Mel Frequency Cepstral Coefficients which encodes the power spectrum of a sound. Ch 10 is about Computer Vision and Pattern Recognition, starting by basic image processing (let us work on beautiful Lenna, again) and how to deal with noise and then moves to pattern recognition. Ch 11 starts with a list of good reason on why Dimensionality Reduction is a useful tool while before we tried to use every additional feature that could help our machine learner. Redundant features, filtering, correlation, mutual information, wrappers… is all about feaure extraction and principal component analysis (PCA), limitations and LDA (linear discriminant analysis) as well as multidimensional scaling. The closing Ch 12 encourages to use the learned methods and tools to work with even bigger data, example, through the use of ‘tasks’, partial results, data analysis; the authors invest carefully this knowledge helping us to construct Machines and the use of Amazon Web Services. Python packages for Amazon Linux (to be used on the Amazon elastic compute cloud!) And as a final demostration of how good are these 2 teachers, they let us know where to look and find for more on Machine Learning. Give yourself a time, get this book, study and use it with your data: suddenly many things that were obscure or abstruse, will became clearer. This books is a investment which worths every penny.
http://www.packtpub.com/building-mach...
The authors assume we hve already installed at least a Python 2.7 version in our machine. Then, in chapter 1 (chapter, hereafter Ch) they promptly explain the main lines of the book, what we could learn from it and how/when/where to apply this knowledge. Then they focus in the Python tools/libraries needed:, Numpy, Scipy, Matplotlib and Scikit-learn, and how to use them to get the most of our data. They reveal the power of those libraries: a particularly nice example is given at the very beginning when they compare runtime behaviours. They end the first chapter with a full fledged example: goal, methods, needed tools and vocabulary with words as overfitting, test and training, in the context of machine learning.
In Ch 2 they move into real life examples and the use of ‘external’ (though, classical) datasets and the notion of ‘classification’ first through the visualization of the dataset and then through coding for the extraction of features, construction of models (able to be used with different, new datasets), their evaluation, and more complex classifiers. Ch 3 is about Clustering (Flat and Hierarchical) and the finding of Related Posts in, say,a blog website, through the search and pre processing of similarities between common words using the notion of words stem and how to deal withy the related noise. In Ch 4 the authors start the discussion by taking this book as an example on how to classify the book (machine learning or python?) expanding from clustering to the notion of topics. Cleverly, they introduce a couple of concepts and the acronym LDA both as an exercise as well as important definitions: we learn on every page! In Ch 5 they are back into classification and teach us how to detect poor answers example, through the preselection and processing of attributes and defining what is a good answer and how to deal with variances and bias, accuracy and precision. Ch 6 is about …. Sentiments and (example) Twitter opinion mining, critical for commerce and politics through the use of the Naïve Bayes classifier (and Theorem) one of “the most elegant machine learning algorithms”. In Ch they give us some Recommendations on Regression plus the notion of hyperparameters. Ch 8 provides even better ideas and methods for Regression and Recommendations and what happens when there are not numerical ratings, introducing the Basket Analysis method (example, Amazon “customers who bought this also bought…”) applicable when we want to to do a recommendation example, based on predictors and association rules. Ch 9 on Music Genre Classification, not by, say, titles, but actually inspecting the data, example, the spectrograms and decomposition on wave components. Hardly could youu imagine it more technical, inspirating and useful! This allows to go into more specifics as the Mel Frequency Cepstral Coefficients which encodes the power spectrum of a sound. Ch 10 is about Computer Vision and Pattern Recognition, starting by basic image processing (let us work on beautiful Lenna, again) and how to deal with noise and then moves to pattern recognition. Ch 11 starts with a list of good reason on why Dimensionality Reduction is a useful tool while before we tried to use every additional feature that could help our machine learner. Redundant features, filtering, correlation, mutual information, wrappers… is all about feaure extraction and principal component analysis (PCA), limitations and LDA (linear discriminant analysis) as well as multidimensional scaling. The closing Ch 12 encourages to use the learned methods and tools to work with even bigger data, example, through the use of ‘tasks’, partial results, data analysis; the authors invest carefully this knowledge helping us to construct Machines and the use of Amazon Web Services. Python packages for Amazon Linux (to be used on the Amazon elastic compute cloud!) And as a final demostration of how good are these 2 teachers, they let us know where to look and find for more on Machine Learning. Give yourself a time, get this book, study and use it with your data: suddenly many things that were obscure or abstruse, will became clearer. This books is a investment which worths every penny.
September 30, 2013
Machine learning is an intricate philosophy and it involves lot of mathematical complexities to bring it into a practice of data analysis. This book simply eradicate those intricacies of programming and implementation of machine learning algorithms. In all, it makes machine learning code pretty simple. Understanding "WHAT" is machine learning is not the purpose of this book. However, this book is designed around the concept "HOW" to implement machine learning algorithms. I would like to add here that it is not only explain you "HOW" to program the algorithms but it also helps you to think "HOW BEST" we can program it. Let me start with some + and few - of the books. But before that remember, as title clarifies, this book is all around (hovers around) Python implementation of machine learning i.e. SCIKIT-LEARN libraries, Scipy and NUMPy. That's the boundary.
+
1. Very clear and precise declaration from Author that this book is more about implementation of ML than Concept.
2. It starts with teaching very basic of data analysis of preprocessing and cleaning up the data along with implementation of Array, indexes, Vector and Matrices using python libraries. This helps reader to make aware about WHAT basics they should build before getting into more complex problems of machine learning. I really liked the "tiny" machine learning program. It's like writing "Hello Word" in any other programming book.
3. Beauty is that it takes you slowly into the implementation of classification problem, Text data processing, Clustering, Regression and sentiment analysis.
4. Though the breadth of topics is vast but it touches every small corner of related topic. For example: When explaining text comparison method it explains how STOP WORD can be done? how to implement TF-IDF for meaningful text comparison? etc.
5. I had big time difficulties in understanding correlation and regression. but explanations of supported SCIKIT libraries made it pretty simple.
6. I really enjoyed the chapter for implementation of text (post) data comparison and clustering.
7. Big data analysis using JUG came as surprise to me when i was about to complete the reading. It is really interesting to compare this topic with Map reduce implementation. My work is still in progress on this...
8. overall this book covers almost everything that PYTHON can cover for you in data analysis and machine learning.
-
1. You should know ML concepts in advance. This is not the book to start ML learning. Obviously, It's already proclaim that it is programming ML in python.
2. Don't expect in details explanation of any algorithm. Like, when it says "TF-IDF" it just explain in a paragraph what TF-IDF is. And, it implementation in Scikit-learn libraries.
3. You mush have moderate level of understanding on python. If you are not at all familiar with PYTHON then spend some time on python primitive data types and programmability before you start this book.
4..... yeah that's it. I don't have any more points to mention as negative.
Overall, very good book when you have some knowledge on ML (and its algorithms) and Python. But you don't know how to implement these concept in data analysis. Then this is the book. Go get it!
+
1. Very clear and precise declaration from Author that this book is more about implementation of ML than Concept.
2. It starts with teaching very basic of data analysis of preprocessing and cleaning up the data along with implementation of Array, indexes, Vector and Matrices using python libraries. This helps reader to make aware about WHAT basics they should build before getting into more complex problems of machine learning. I really liked the "tiny" machine learning program. It's like writing "Hello Word" in any other programming book.
3. Beauty is that it takes you slowly into the implementation of classification problem, Text data processing, Clustering, Regression and sentiment analysis.
4. Though the breadth of topics is vast but it touches every small corner of related topic. For example: When explaining text comparison method it explains how STOP WORD can be done? how to implement TF-IDF for meaningful text comparison? etc.
5. I had big time difficulties in understanding correlation and regression. but explanations of supported SCIKIT libraries made it pretty simple.
6. I really enjoyed the chapter for implementation of text (post) data comparison and clustering.
7. Big data analysis using JUG came as surprise to me when i was about to complete the reading. It is really interesting to compare this topic with Map reduce implementation. My work is still in progress on this...
8. overall this book covers almost everything that PYTHON can cover for you in data analysis and machine learning.
-
1. You should know ML concepts in advance. This is not the book to start ML learning. Obviously, It's already proclaim that it is programming ML in python.
2. Don't expect in details explanation of any algorithm. Like, when it says "TF-IDF" it just explain in a paragraph what TF-IDF is. And, it implementation in Scikit-learn libraries.
3. You mush have moderate level of understanding on python. If you are not at all familiar with PYTHON then spend some time on python primitive data types and programmability before you start this book.
4..... yeah that's it. I don't have any more points to mention as negative.
Overall, very good book when you have some knowledge on ML (and its algorithms) and Python. But you don't know how to implement these concept in data analysis. Then this is the book. Go get it!
December 10, 2014
Quite an impressive book. I’m not sure what has amazed me most: the examples, so real and complex but still, easy to follow, or the easiness through which the authors introduce the reader to many different machine learning techniques without throwing at him scaring math formulas. Certainly both things. And not only those.
As clearly stated in the preface, this book it’s not about scikit-learn. It does not expect the reader to master it. Similarly, it does not aim to teach him the library from install to Z. The book is about Python and machine learning. Many examples are, indeed, pure Python.
Still, scikit-learn is used. Sometimes alone, often in combination with other widely known libraries, such as Numpy, Scipy, nltk and matplotlib. While scikit-learn and nltk are not taken from granted, whoever is holding the book on his hands is expected to have been exposed already to both the numerical libraries and to matplotlib. Methods such as norm() and rvs() are not explained and if you are not familiar with them, get ready to open both the reference and a terminal with the Python interpret.
Not only do the authors rely on those must know libraries. They also show the readers some less known gem, such as gensim (a real beauty that doesn’t ship with scikit-learn).
As I was saying at the beginning, the examples are very interesting. They are real world challenges, clearly explained step by step with many accompanying charts and schemas. The source code that comes with the book has some snippet that you really wanna copy somewhere safe.
I have particularly enjoyed the idea behind the book itself: introduce the reader to machine learning without exposing him to the mathematical details behind it. The concepts are very well explained and the few formulas you will see won’t get you screaming terrified. I pretty much agree with this approach: we don’t need to know how an engine works to drive a car.
I do absolutely recommend this little masterpiece to any Python programmer who wanna get started with machine learning. This book really get you curious about the subject. Isn’t it what all great books do?
As usual, you can find more reviews on my personal blog: http://books.lostinmalloc.com. Feel free to pass by and share your thoughts!
As clearly stated in the preface, this book it’s not about scikit-learn. It does not expect the reader to master it. Similarly, it does not aim to teach him the library from install to Z. The book is about Python and machine learning. Many examples are, indeed, pure Python.
Still, scikit-learn is used. Sometimes alone, often in combination with other widely known libraries, such as Numpy, Scipy, nltk and matplotlib. While scikit-learn and nltk are not taken from granted, whoever is holding the book on his hands is expected to have been exposed already to both the numerical libraries and to matplotlib. Methods such as norm() and rvs() are not explained and if you are not familiar with them, get ready to open both the reference and a terminal with the Python interpret.
Not only do the authors rely on those must know libraries. They also show the readers some less known gem, such as gensim (a real beauty that doesn’t ship with scikit-learn).
As I was saying at the beginning, the examples are very interesting. They are real world challenges, clearly explained step by step with many accompanying charts and schemas. The source code that comes with the book has some snippet that you really wanna copy somewhere safe.
I have particularly enjoyed the idea behind the book itself: introduce the reader to machine learning without exposing him to the mathematical details behind it. The concepts are very well explained and the few formulas you will see won’t get you screaming terrified. I pretty much agree with this approach: we don’t need to know how an engine works to drive a car.
I do absolutely recommend this little masterpiece to any Python programmer who wanna get started with machine learning. This book really get you curious about the subject. Isn’t it what all great books do?
As usual, you can find more reviews on my personal blog: http://books.lostinmalloc.com. Feel free to pass by and share your thoughts!
December 10, 2014
Quite an impressive book. I’m not sure what has amazed me most: the examples, so real and complex but still, easy to follow, or the easiness through which the authors introduce the reader to many different machine learning techniques without throwing at him scaring math formulas. Certainly both things. And not only those.
As clearly stated in the preface, this book it’s not about scikit-learn. It does not expect the reader to master it. Similarly, it does not aim to teach him the library from install to Z. The book is about Python and machine learning. Many examples are, indeed, pure Python.
Still, scikit-learn is used. Sometimes alone, often in combination with other widely known libraries, such as Numpy, Scipy, nltk and matplotlib. While scikit-learn and nltk are not taken from granted, whoever is holding the book on his hands is expected to have been exposed already to both the numerical libraries and to matplotlib. Methods such as norm() and rvs() are not explained and if you are not familiar with them, get ready to open both the reference and a terminal with the Python interpret.
Not only do the authors rely on those must know libraries. They also show the readers some less known gem, such as gensim (a real beauty that doesn’t ship with scikit-learn).
As I was saying at the beginning, the examples are very interesting. They are real world challenges, clearly explained step by step with many accompanying charts and schemas. The source code that comes with the book has some snippet that you really wanna copy somewhere safe.
I have particularly enjoyed the idea behind the book itself: introduce the reader to machine learning without exposing him to the mathematical details behind it. The concepts are very well explained and the few formulas you will see won’t get you screaming terrified. I pretty much agree with this approach: we don’t need to know how an engine works to drive a car.
I do absolutely recommend this little masterpiece to any Python programmer who wanna get started with machine learning. This book really get you curious about the subject. Isn’t it what all great books do?
As usual, you can find more reviews on my personal blog: http://books.lostinmalloc.com. Feel free to pass by and share your thoughts!
As clearly stated in the preface, this book it’s not about scikit-learn. It does not expect the reader to master it. Similarly, it does not aim to teach him the library from install to Z. The book is about Python and machine learning. Many examples are, indeed, pure Python.
Still, scikit-learn is used. Sometimes alone, often in combination with other widely known libraries, such as Numpy, Scipy, nltk and matplotlib. While scikit-learn and nltk are not taken from granted, whoever is holding the book on his hands is expected to have been exposed already to both the numerical libraries and to matplotlib. Methods such as norm() and rvs() are not explained and if you are not familiar with them, get ready to open both the reference and a terminal with the Python interpret.
Not only do the authors rely on those must know libraries. They also show the readers some less known gem, such as gensim (a real beauty that doesn’t ship with scikit-learn).
As I was saying at the beginning, the examples are very interesting. They are real world challenges, clearly explained step by step with many accompanying charts and schemas. The source code that comes with the book has some snippet that you really wanna copy somewhere safe.
I have particularly enjoyed the idea behind the book itself: introduce the reader to machine learning without exposing him to the mathematical details behind it. The concepts are very well explained and the few formulas you will see won’t get you screaming terrified. I pretty much agree with this approach: we don’t need to know how an engine works to drive a car.
I do absolutely recommend this little masterpiece to any Python programmer who wanna get started with machine learning. This book really get you curious about the subject. Isn’t it what all great books do?
As usual, you can find more reviews on my personal blog: http://books.lostinmalloc.com. Feel free to pass by and share your thoughts!
March 19, 2015
This is better than Doing Data Science, but still a bit cursory and casual for my tastes. It's a decent overview of some machine learning techniques, but I didn't find the Python implementations particularly informative, really.
January 4, 2016
Excellent tutorial book with many examples and fairly clear explanations. Note that there are some steps missing from the sample code in the book, and a few syntax typos, so be prepared to use your Python debugging skills to figure out what's going wrong with the examples. Other than that annoyance, this is a very good book on the subject.
February 23, 2014
A good book with some serious errors.
June 6, 2015
I was a technical reviewer on this one, so I'm biased. But a really good read, good information.
September 10, 2015
I found the book very basic and rather superficial take on machine learning with quite a few typos and errors. Aimed for entry level.
Displaying 1 - 13 of 13 reviews