For the developers among you looking to take your knowledge of AI and machine learning to the next level of expertise, a new paperback edition of the hardback classic by Brian D. Ripley is scheduled for released in a few days (on the 10th of January to be precise). With an extra 22% pre-order discount from Amazon U.S (or U.K), it’s now particularly affordable to dig deeper into the subject!
Increasingly, games are using pattern recognition and data-mining technology for things like animation, extracting trends from sales statistics or finding patterns in play test data. But few developers have the necessary understanding in statistical machine learning required to find useful patterns quickly amidst all the data.
If you’re keen on developing an expertize of these techniques, then this is the book to get. The original hard-back edition of Ripley’s Pattern Recognition and Neural Networks has been praised for its balance of statistical decision theory with real-world examples. (Though it’s not intended for beginners.)
Here’s the official blurb written by the publisher for the hardback edition:
“This book uses tools from statistical decision theory and computational learning theory to create a rigorous foundation for the theory of neural networks. On the theoretical side, Pattern Recognition and Neural Networks emphasizes probability and statistics. Almost all the results have proofs that are often original. On the application side, the emphasis is on pattern recognition. Most of the examples are from real world problems.
In addition to the more common types of networks, the book has chapters on decision trees and belief networks from the machine-learning field. This book is intended for use in graduate courses that teach statistics and engineering. A strong background in statistics is needed to fully appreciate the theoretical developments and proofs. However, undergraduate-level linear algebra, calculus, and probability knowledge is sufficient to follow the book.”
About the Book
The hardcover edition of the book has been cited by over one hundred other books, and referenced by a large corpus of white papers on the subject. But what makes this book special, compared to the thorough academic publication by Christopher M. Bishop, Neural Networks for Pattern Recognition, is that it’s much more accessible thanks to the many real-world examples.
Of course, the book is not for those starting out in mathematics or computer science. It requires a certain amount of time to pick up some of the mathematical statistics in the book, so it’s helps if you have the background. But it can be a rewarding process, like any learning curve!
Table of Content
Introduction and Examples
Statistical Decision Theory
Linear Discriminant Analysis
Feed-forward Neural Networks
Finding Good Pattern Features
A Statistical Sidelines
Have you read this book? What do you think?