Top positive review
2 people found this helpful
on February 12, 2010
This is the textbook I'm using for an undergraduate machine learning course, and so far it has been very enjoyable. There are plenty of exercises in each chapter, from simple "derive a formula for ..." to more in depth problems, and several of the problems have solutions. One thing I've found really useful is how well referenced the book is within itself. It'll say something like "Recall that, if we assume a squared loss function, then the optimal prediction, for a new value of x, will be given by the conditional mean of the target variable". In the margin, it then has in red the text "Section 1.5.5" pointing you to where we learned this bit of trivia. Formatting is also well done, charts are colourful and seem to get the point across well. I find that I learn much easier by being shown a picture/graph of what we want to achieve, and then have it described, and finally being given the equations for solving this (rather than just being given the equations), which this book does well. It has taken a fair amount of work to get through though, and so I wouldn't say it's an easy textbook by any means (I mean, come on, we're teaching computers how to think, that can't be easy).
My one complaint is that I wish that they had a chapter/appendix with a bit of a stats refresher, because the last stats course I took was over a year ago, and so this textbook took me a little bit to get into for lack of knowing what some of the early terms meant.