Machine Learning for Hackers

Drew Conway and John Myles White

Published by






Reviewed by

Nilesh Thatte MBCS


9 out of 10

'Machine Learning for Hackers' is hands-on introductory book about machine learning using statistical tools, specifically R. Authors Drew Conway and John Myles White have years of experience in statistical analysis and developing tools. They have written this book for people with strong programming skills. 

The book provides a fast track for data crunchers who would like to start their journey into machine learning. It explains various algorithms enabling the reader to understand the working of these methods and their pitfalls.

The first quarter of the book is a gentle introduction to R and its toolset, mainly focusing on visualisation of data for analysis. This is useful starter for readers with no background in R, though at times they may need to refer to R manuals or another book on R.

Each chapter in the rest of the book focuses on a specific problem like classification, predictions, regression and recommendations.

The book is full of good ideas and practical case studies. Google Social Graph API was used to illustrate graphical thinking; while it is not supported any more, it still is a useful learning experience and provides a foundation to explore this subject area further.

The last chapter introduces SVM (Support Vector Machine) technique for non-linear classification and compares it with other algorithms. A more detailed coverage on this difficult and interesting subject area could have been useful in this book.

The R script and data is available for download, which is a standard nowadays.

I will give book 9 out of 10, and will recommend this book to anyone with interest in machine learning with a statistics background.

Further information: O'Reilly

August 2012