[Noisebridge-discuss] Machine Learning seminars?

Meredith L. Patterson mlp at thesmartpolitenerd.com
Wed Dec 10 21:44:15 UTC 2008


Josh Myer wrote:
> 3.) Data Mining: Practical Machine Learning Tools and Techniques by 
>      Witten and Frank
> 
> http://www.amazon.com/Data-Mining-Practical-Techniques-Management/dp/0120884070/ref=pd_bbs_sr_1?ie=UTF8&s=books&qid=1228940709&sr=8-1
> 
> I rather like this book, but it's a little heavy for introductions.

Witten and Frank has the added bonus that it works hand in hand with
Weka, which is a solid, mature toolkit that can be used standalone or as
Java libraries. A++ would classify again.

> I'm also open to suggestions from others; my background is mostly
> second-hand from practitioners.

There's one other book that I used as an intro when I started working
with machine learning, but I can't remember the name off the top of my
head. I'll check my shelves when I get home.

Really, though, once you get past Witten and Frank, you have to leave
the buffet and pick an area to focus on (Bayesian analysis, perceptrons,
support vector machines, decision trees, text classification, choose
your poison).

_Pattern Classification_ by Duda, Hart and Stork is fantastically
thorough, but heavy on the math; I'd say only about a quarter of it is
accessible to someone who isn't very, very comfortable with linear
algebra. (It's possible to view a lot of what the book covers in
geometric terms instead, but the book doesn't provide a lot of material
to make that easy.)

--mlp



More information about the Noisebridge-discuss mailing list