|
As there is already a question about Good Freely Available Textbooks on Machine Learning, I wanted to ask the same sort of question but without the restriction on free. As much as I enjoy freedom of knowledge, in a smaller community some of the best texts simply cannot be found for free. If anyone knows of any machine learning textbooks, general or task-specific, that truly helped them understand what they know today, I ask that you please mention and link them here. Also, if any tutorial documents or research papers have been particularly clear, I ask you also mention them. The text I wish to recommend is unfortunately not yet available except to University of California, Berkeley students as it is still in draft form. It will be called An Introduction to Probabilistic Graphical Models when it is available and is purchasable by students of CS281a. Unfortunately, Professor Jordan has asked draft copies not be distributed. |
|
I suspect that it will be hard to beat Michael Jordan's book if you want something with strong statistical foundations. I say that based solely on the author. Coming from a statistics background, I have found most ML books to be very unsatisfying reading. Michael Jordan's writing is a stark exception. I would recommend reading statistics books as a complement to ML papers/books. For instance, Gelman et al's Bayesian Data Analysis and Robert's Bayesian Choice are both excellent Bayesian statistics books. The first is more applied, focusing careful analysis of small scale problems, and the latter is more theoretical, focusing on decision theory. While the types of applications these books study are not central to machine learning, the language and framing is useful, particularly as the machine learning literature moves towards the statistical approach. |
|
if you want a good graphical models book that is actually available, i'd go with daphne koller's: http://www.amazon.com/Probabilistic-Graphical-Models-Principles-Computation/dp/0262013193/ref=ntt_at_ep_dpi_1 I haven't read mike jordan's book, only some of his papers, and I think his treatments are good if you already know the material well, but it's sometimes hard to use his material to learn something new. |
|
Look to following blog posts (first, second, and may be this one) - they contains pointers onto big number of books, related to machine learning, and related topics I forgot to add link to original article - http://measuringmeasures.blogspot.com/2010/01/learning-about-statistical-learning.html - in comments, there are a lot of links to books. And I also remember, that discussion on Hacker News also had a lot of links - http://news.ycombinator.com/item?id=1055042
(Jul 03 '10 at 05:19)
Alex Ott
Unfortunately the link that u mentioned again, does not work :(
(Nov 03 '12 at 09:10)
Lancelot
|
Has anyone read Foundations of Machine Learning by Mohri et al, https://mitpress.mit.edu/books/foundations-machine-learning-0 ?
I'm looking for a good solid introduction to the theory side of ML.