Support Vector Machines....this goes out to Baltazar and possibly FDAX

Nonius · Post by **Nonius** » Thu Dec 09, 2004 1:37 pm

I seem to recall FDAX mumbling about this about 1000 years ago, and, also, I think Baltazar may know a fair amount about this. Some stuff crossed my desk on the subject, and I wanted to get:

1. some simple tutorial/explanations here.

2. references of good articles, papers to read on the subject.

FDAXHunter · Post by **FDAXHunter** » Fri Dec 10, 2004 12:50 pm

Sorry, I know I'm late. Forgive me:

Support Vector Machines by Marti A. Hearst from, yes, Berkeley! Very good. Probably the best introduction.

Support Vector Machines a sort of step-by-step walkthrough

Baltazar · Post by **Baltazar** » Fri Dec 10, 2004 1:34 pm

well i'm later than FDAX but here they are:

a web site on the subject: http://www.kernel-machines.org/

tutorials:the one fdax gave is good.

you may find these papers interesting

http://www.kernel-machines.org/papers/upload_13550_kmrev.ps

http://www.kernel-machines.org/papers/Burges98.ps.gz

http://www.kernel-machines.org/papers/smosch97.ps.gz

http://www.kernel-machines.org/papers/tr-30-1998.ps.gz

http://www-ai.cs.uni-dortmund.de/DOKUMENTE/klinkenberg_joachims_2000a.ps.gz

this one on time serie prediction

http://www.kernel-machines.org/papers/IC.ps.gz

ftp://ftp.esat.kuleuven.ac.be/pub/SISTA/suykens/reports/lssvm_00_79.ps.gz

this one in density estimations

http://www.ai.mit.edu/projects/cbcl/publications/theory-learning.html

Nonius · Post by **Nonius** » Fri Dec 10, 2004 1:40 pm

thanks dudes, I knew you'd come through....

quantie · Post by **quantie** » Fri Dec 10, 2004 6:55 pm

There are two books that I'd recommend duda & hartand the book by Hastie. There is this guy selling on amazon-us who ships all the way from china so you get it for a fraction of the cost!.

Baltazar · Post by **Baltazar** » Fri Dec 10, 2004 7:11 pm

yes good ones.

i think the duda is free on the web somewhere

also of course vapnick's books 'statistical learning theory' and 'the nature of statistical learning theory'. it's kind: the maths in the first (quiet hard) and the philo of it in the second.

dgn2 · Post by **dgn2** » Fri Dec 10, 2004 7:24 pm

duda is also a favourite of mine.

Nonius · Post by **Nonius** » Fri Dec 10, 2004 9:11 pm

(empty post)

dgn2 · Post by **dgn2** » Sat Dec 11, 2004 11:42 am

Nonius, I have been running backtests on my workstation for a couple of weeks, trying to uncover anything I may have missed. I didn't have much time to work on my models over the last couple months until recently and I had nearly forgotten how dirty LIFFE tick data can be. The sheer volume of data for stuff like EURIBOR is getting unmanageable...my MySQL queries are taking too long...I had this idea that I could infer an intraday EUR swap curve from the EURIBOR and 5y & 10y SwapNotes for use in some relative value indicators, but the 5y SwapNote data is of such low quality that I am having second thoughts about the validity of this approach...building a system for cleaning this data is getting monotonous...plus I need a cluster or something .

Nonius · Post by **Nonius** » Sat Dec 11, 2004 1:39 pm

I'm interested in hearing more of your rants on the intraday data stuff.....