|
Hi All, I've been asked for recommendations for books or tutorials specifically relating to introductory / 101 level "Math for Machine Learning". Any pointers appreciated. Christopher |
|
See the list of books from this blog and pick a book.. or more. Great List. Thanks for pointer.
(Jan 31 '11 at 13:53)
communicating
|
|
I think chapter 29 in Bayesian Reasoning and Machine Learning (Barber, 2011) is a quite good overview. The book is getting released in June I think, but there is a pdf online. Very Good Overview Chapter, I think it's part of what he needs, thanks! :)
(Jan 31 '11 at 13:52)
communicating
|
|
I'll recommend you 2 books, first is the CHapter 2 of Bishop's Machine Learning. The second would be the MAtrix Cookbook, it has most of the identities and derivations you would ever need for most of the optimization problems, which is one of the weak point of most people going into machine learning, too much probability but little Matrix Math. Matrix CookBook is here |
|
Most of the machine learning books I've read are pretty clear on the ML math parts. My big problem was I had forgotten nuances of Linear Algebra and Calculus that were getting in my way of understanding some details. This site has a lot of free books that filled in what I had forgotten: http://www.freebookcentre.net/SpecialCat/Free-Mathematics-Books-Download.html. I found the Linear Algebra Complete book the most helpful for me. A similar place is http://www.sosmath.com/, but somewhat scarcer that your site.
(Feb 01 '11 at 10:14)
Lucian Sasu
|
|
Hi Christopher I recommend you to read chapter 2 of the book "Advanced lectures on machine learning: ML Summer Schools 2003". Arya |
|
There was reading group at Wisconsin last year with almost the same title, and the web page has some good resources: http://pages.cs.wisc.edu/~andrzeje/lmml.html |
|
Once you get past the basics ('101' as you put it), you might find it convenient to study a couple of math topics in more depth to help yourself become more comfortable communicating in this language -- eg: linear algebra, probability & statistics, graph theory, numerical analysis (to name a few). Most books, wikipedia pages, etc, will have a tendency to stick to a fairly typical set of notational styles. However, once you start getting into reading research papers or other more fluid outlets of technical information, it can be rather irritating to figure out what the author means by the notational conventions they're using -- particularly if your exposure/experience with various notational conventions and topics in math is sketchy. -Brian
This answer is marked "community wiki".
|
