Glenn said
Collective Intelligence is a big topic so the title of this book is very ambitious. The author is too. He sets out to explain all things collective intelligence circa 2009.
Like any good text book, the material is presented in an iterative and incremental way. High level concepts are introduced first. Popular features such as tag clouds and recommendation engines are defined. How collective intelligence manifests in the GUI and screen flow of various popular web properties is illustrated. Various data structures and algorithms are classified. Simple examples of some of the underlying math is worked out. Open source libraries are recommended and coding examples presented. Topics are introduced simply at first and revisited in later chapters with ever increasing amounts of detail.
However, there's a lot of math in the topic of collective intelligence and this is not a math book. Neither does it take a "heads up" or "for dummies" approach. Instead, it aims somewhere in the middle which gives the book a bit of an identity crisis.
If you're a script jock or a web master whose level of technical competence doesn't extend too far past embedding a youtube video or paypal "buy now" button on your pages, then the most that you are going to get out of this book is just the high level introductory material.
The formally trained software engineer can treat this book as a great introductory survey level course on the subject that does attempt to peal back some of the layers of complexity. It is neither definitive nor canonical so you will be doing more research and studying before you deploy any sophisticated collective intelligence in your applications.
Jun Zhang said
The publisher has fixed the most errata listed on O'reilly website in the edition I got from Amazon. By running sample codes provided in this book, I
gained invaluable first-hand experiences on various algorithms. Also Chapter 12 on
algorithm summary is nice: author listed strengths and weaknesses for each
algorithm mentioned in previous chapters. I recommend this book for any person
who is interested in machine learning and data mining.
Developpez.com writers said
In my opinion, this book is intended for people with good knowledge in Business Intelligence as well as Java to understand the examples provided, but this is a practical book for the application of collective intelligence in a web applications.
Being divided into three main parts, the book introduces us, through the 1st, 2nd and 3rd chapters, to the field of collective intelligence by giving us an overview of the architecture necessary to integrate collective intelligence in an application.
The 2nd part of the book deals with the intelligence derived from the collection of data. Through the 7th chapter, the reader is introduced to the process of data mining and its different types of algorithms. WEKA tool was used, an open source tool for data mining whose use will be more explicit in Chapter 9 through examples of clustering.
The last part deals with the application of intelligence in an application.
The strengths of the book lie in the methodology adopted by the author to explain how to apply collective intelligence in a web application. Thus, the book explains, through many examples, how to apply collective intelligence, by enumerating its profits in a web application. Each concept is illustrated by concrete examples. The author cites several notable sites using the concept of collective intelligence, something that may give more visibility to readers who wish to implement this concept. One last thing I greatly appreciated in this book is that it brings us from the world of abstract art to a world of practice in predicting the need for Internet users.
Some technical aspects presented by the book may cause the loss of the thread during playback, such as the presence of several class diagrams and Java code. I would have preferred that the book does not address the conceptual aspects of things through the technique. The same problem is encountered by the presence of some mathematical formulas, I would have preferred that the author offers easy ways to implement in order to achieve useful results if it is necessary to go by numbers (calculation of the correlation , similarity, ...).
What I did not like in the methodology of applying data mining in this book is the use of an API directly from Java WEKA instead of using WEKA tool . This is to avoid reinventing the wheel. Also, the author is expected to explain how to select attributes predictive for the construction of a predictive model.
Jan Krivocheia said
Collective Intelligence in Action delivers an in depth account and perspective on scalable data mining techniques. This book is an excellent resource to engineers, architects, and anyone who wants to know more about the topic of collective intelligence. I really recommend this book to anyone who is interested in developing applications with personalization.
Kevin Hoffman said
I've been interested in CI for quite some time now and I've read a lot of books on it, including the other Python book. CI involves a lot of math that many people, even programmers, probably aren't familiar with (or, like me, learned it in college and then forgot it all).
This book presents the material in a clear, concise, practical manner using examples and showing how the concepts work before getting into the code. This makes the code samples far easier to understand. Plus, with the code samples being in Java, all the samples in this book are ready to plug right into whatever back-end enterprise project you're working on.
If you are going to decide on just one CI book to pick up... pick this one, it is by far the best collective intelligence book available.
Comments