Machine Learning in Action

I am new to Machine Learning and I found the book a very good hands-on introduction on the subject. The author takes 8 of the Top 10 algorithms in Machine Learning (based on a 2007 survey paper) and implements them in Python. Other reviewers have pointed out that the theoretical explanations and code quality were somewhat lacking, and thats true. However, even though Python is an extremely readable language, machine learning algorithms are (generally) hard, and I found that it helped to understand them better if I typed them out myself, copying/copy-pasting and restructuring the code as I went, and experimenting with the contents of the intermediate data structures in the REPL. Also, once you have a general idea of how it works, it becomes easier to parse the math in the paper on which the algorithm is based.I still don't completely understand all the implementations, but the book did give me some intuition about how to choose the right algorithm for a given problem. I believe that is also important since ML practitioners often use third party algorithms rather than code everything up from scratch. Of course, for the times you do need to code it up from scratch, you can get some valuable insights about machine learning algorithm design from the style adopted in the book - start small, visualize in 2D/3D for insights, then generalize to higher dimensions. The examples cover a wide range, from dating sites to semiconductor plants, so you get a feel for all the different places these algorithms can be applied.In short, if you want to "Just do ML", ie, quickly get started and pick up anything else you need along the way, then this book may be for you.

This book is more details oriented and thorough than the other books I have read so far.Chapter 1: How do you find K in KNN? is it the square root of the population size?How do you find errors in KNN? The author has discussed only the accuracy metrics.Is there any room for cost adjustment (like weighted distance measurement)?Finally, I believe, KNN has a better application than handwriting identification because that works better with NN.Chapter 5: While estimating the loss, simply used simple weight differences. A cross-entropy function could have been discussed.

Great Book for machine learning and Python lover, I am sure it is a great book for people who are not familiar with matlab, this may be the fast way to get yourself to do some really work rather than keep reading a lot of papers or pseudo codes.It is not very difficult to read and practice, sometimes you may think of some better ideas from the book.Also the idea of machine learning methods could also be helpful when you use other programming language to do some designs.

I teach data science, and have read many many many books on the subject. Of them all, I feel Machine Learning In Action is the best because it is the most clear, has the best example code, and covers the most topics. I recommend this book for beginning data scientists, and advanced alike. The example code and data is also on github. If you are a python user, this book is a MUST.

Looking at many good reviews on amazon, I decided to purchase this book. It's a decent book, but IMO it has been edited poorly and the code has not been tested properly.The introduction chapter got me really excited, just like other Manning's "in Action" books do. But once I started executing the code in chapter 2 "Classifying with k-nearest neighbors" I realized that the code had bugs. Though I could figure out what's wrong and fix the bugs, I did not expect this from Manning, after having read some of their excellent books like ( The Quick Python Book, Second Edition , Spring in Action  and  Hadoop in Action ).Moreover the book has some introduction to python and numpy in appendix A. I believe the author could have pointed the reader elsewhere for learning python and those pages could have been used to explain more of numpy and matplotlib, which the author uses freely without any explanation in the text. (Yup, be ready to read some online numpy and matplotlib tutorials and documentation.)If you don't know python, then you can do what I did: read  The Quick Python Book, Second Edition  and then attempt this book.The figures in the book are not in color so you need to execute the code to understand what the author is telling. It forces you to actually run the code, which is good, but you can't read this book without a computer in front of you.Finally, I am a big believer in following the conventions of a language. I would have been really happy had the author followed PEP8 ([...]), because along with learning machine learning, you could have learnt some good python coding practices.

No lo use como tal, pero me pareció buen libro

A really good book that introduces ML algorithms. Many common ML algorithms are introduced and implemented. The book may appear a bit complex for someone who just started machine learning. Mix the contents of this book with some good courses online, and you are good to go.The book is not for the impatient or faint hearts. The book shows actual implementation of various ML algorithms in Python using NumPy library.

is it too much for 2 stars?the codes have too much bugs

Testo introduttivo ai classici e maggiormente noti temi di Machine Leraning, declinati in una variante particolarmente orientata all'applicazione ( in Python ).Corredato da numerosi esempi di codice più o meno "completi" fanno di questo testo una reference di alto livello per chi volesse avvicinarsi per la prima volta al mondo del Machine Learning utilizzando Python e le librerie scientifiche che mette a disposizione.Consigliato!

Der Autor ist gelernter Elektronikingenieur. Er hat ein paar Jahre bei Intel gearbeitet. 2008 hat er sich - ohne zu inskribieren - erstmals in eine Statistikvorlesung gesetzt. Laut Buchrücken hat er bereits in zahlreichen akad. Journalen publiziert. Ich konnte kein ihm zuordenbares Werk ausfinding machen. Eine Anfrage im Verlagsforum blieb bisher unbeantwortet.Ich vermute vielmehr, dass Harrington seine eigenen Lernschritte und praktischen Experimente zu einem Buch verarbeitet hat. Er überfordert den Leser jedenfalls nicht mit hochgestochener Theorie und Mathematik. Persönlich habe ich mich manchmal eher unterfordert gefühlt. Das Buch hat allerdings auch nur den Anspruch einer Einführung.Die Auswahl der Algorithmen orientiert sich an den "Top 10 Algorithms in Datamining" ([1]). Er präsentiert im Buch davon acht. Die Auswahl ist plausibel. Sein eigener Beitrag ist die Programmierung von einfachen Varianten in Python. Nachdem ich noch nie in Python programmiert habe, kann ich die Qualität des Kodes nicht beurteilen. In Rezensionen auf Amazon.com wird er relativ heftig kritisiert. Auf der Verlagsseite gibt es jedenfalls einige errata. Kode ordentlich zu testen scheint heute bei Buchpublikationen nicht mehr üblich zu sein.Die für die Algorithmen verwendeten Beispielanwendungen sind - so wie das gesamte Buch - durchwachsen. Bei 2 Algorithmen verwendet der Autor die Überlebenschance bei Pferdekoliken. Er räumt ein, dass er von Pferden keine Ahnung hat und daher die Ergebnisse nicht beurteilen kann. Einige konstruierte Beispiele sind eher kurios. Z.B. der Zusammenhang zwischen IQ und der Anzahl der Gänge bei einem Fahrrad. Sehr gut hat mir hingegen eine Untersuchung über Abstimmungsmuster von US-Abgeordneten gefallen. Die größte Stärke des Buches ist überhaupt die Auflistung einer Reihe von interessanten Datenquellen. Im Anhang geht er noch auf MapReduce und Hadoop ein. Damit kann man mit von amazon angemieteten Serverfarmen sehr grosse Datenmengen durchackern. Es ist aber mehr eine Werbeeinschaltung für das Hadoop in Action Buch.Das Buch leidet auch etwas an den mässigen Grafiken. Es werden in Scatterplots Punkte aus verschiedenen Gruppen angezeigt. Der jeweilige Algorithmus soll die Gruppen separieren. Es ist aber kaum bis gar nicht erkennbar, zu welcher Gruppe ein Punkt gehört. Eine derartige Darstellung ist nur in Farbe sinnvoll. Gute Bücher (siehe [2]) verwenden dazu auch Farbgrafiken.Prinzipiell finde ich die Betonung des praktischen Aspektes, sich mit Daten die Hände schmutzig zu machen, an diesem Buch sehr sympathisch. Die Ausführung hätte aber in einigen Details wesentlich besser sein können.[1] Xindong Wu et al.: Top 10 Algorithms in Data Mining. Dieser - sehr bekannte Artikel - basiert auf einer Meinungsumfrage unter den Teilnehmern der IEEE International Conference on Datamining, Dec. 2006.[2] Bishop Christoper M.: Pattern Recognition and Machine Learning.