Matlab in Finance

Quote from prophet:

Matlab's average performance was tested to be 5 times better than Mathematical for various numerical operations.
More...

that matches my experience, matlab is noticeably faster for crunching numbers. post wasn't meant as a "mine is bigger than yours" swipe, both are fab tools and either will do more than 99% of us could ever dream up.

mathematica's strength - and this certainly YMMV territory - is that it is much faster in getting from the germination of an idea to full-blown testing of the idea. it's just a terrific development platform. 90% of my time is spent in development, not crunching, so for me this is huge advantage. it is also very very easy to go from testing things out to presentation/documentation, which is a major issue (for me) with Matlab.

some examples of Mathematica based quant offerings...

http://www.ifs.dk/DerivativesExpert/DerivativesExpert.html
http://www.unriskderivatives.com/product/examples.html

a book with lots of goodies..."Modelling Financial Derivatives with Mathematica"...

http://www.ifs.dk/DerivativesExpert/DerivativesExpert.html

again, both platforms are terrific, no complaints about either. there was a time i simulated ridiculously complex systems - we were designing silicon-based retinal workalikes - where Matlabs speed advantage really did matter more than anything else.

if i had to generalize - if all you're doing is grabbing price data and looking for patterns, time series, etc, Matlab is probably a better fit, if you're testing using pricing models, monte carloing, etc, Mathematica will probably be an easier path.

oh...Wolfram has a 64-bit version for 64-bit linux running on Athlon/Opterons. very very nice. :)
 
For those who are interested, it looks like an IB-mathematica interface has been done before.

http://www.wolfram.com/solutions/finance/datafeed.html

One good thing about matlab is that you can "vectorize" your code to work in matrices and vectors and reduce the need for For loops. This has made a HUGE difference in the few projects I have worked on. (maybe as much 100x) Maybe For loops in mathematica don't have the efficiency problems that they do in matlab.
 
If you want mathematica but don't like the price, I believe there is an academic version. It's a cut down version but unless you're planning on winning the Nobel Prize it would probably cater to most traders needs.

Runningbear
 
Quote from damir00:

mathematica's strength - and this certainly YMMV territory - is that it is much faster in getting from the germination of an idea to full-blown testing of the idea. it's just a terrific development platform. 90% of my time is spent in development, not crunching, so for me this is huge advantage. it is also very very easy to go from testing things out to presentation/documentation, which is a major issue (for me) with Matlab. .
More...
Yeah. Speed of prototyping is mainly why we use these tools. Code the performance critical stuff in C anyway.

Matlab really needs some kind of layout editor for arranging editable figures along with text.

I tend to be limited by processing speed. Although I sometimes wish this wasn’t the case, the complexity and constant number crunching has proven more illuminating versus trying to mentally hypothesize what will work. This may go against the grain of advice on ET to "use your brain, not your computer". Actually I do both. I just wouldn't be as successful right now if I had 50% less computationally ability in the last few months.

oh...Wolfram has a 64-bit version for 64-bit linux running on Athlon/Opterons. very very nice. :)
More...
Nice. I wonder it is NUMA-intelligent such that you can allocate your work to particular processors and memory banks. I’m not sure if the Matlab AMD64 version offers that.
 
I have not used it myself, but there is a free matlab clone called Octave that is available. Supposedly you can run matlab m-files on it and the syntax is essentially the same. Maybe someone who knows could confirm this. It is lacking the toolboxes which are available with matlab, though.

http://www.octave.org/
 
Octave is good, but i think it's better suited to those already familiar with Matlab or at least with mathematical programming in general because there is a lot less handholding/docs/etc. Matlab itself derives from open source (ish) roots, so it's kinda neat seeing it all grown up and attracting a second generation competitor from a similar place.
 
Anyone know how much effort it would take a professional programmer to, as Mr. Medved says, "put together the needed connector" between Matlab and Quotetracker? The way I look at it, the ability to hook up to Quotetracker would save me thousands a year (the difference in monthly data fees between something that currently works with the Matlab Datafeed Toolbox, such as Bloomberg, and the cheaper alternatives that work with QuoteTracker, such as eSignal). Anyway, I'd gladly pay a programmer to do this if it wasn't too expensive.


Quote from j_medved:

FYI - We now have a Socket based API for tick by tick data from QT:
http://www.quotetracker.com/help/QTStreamingAPI.shtml

I am not familiar with Matlab, nor do I have the software. Since our API is public and fully documented and quite simple, it would be far easier for Matlab to put together the needed connector, thus in affect adding support for all the datafeeds QuoteTracker supports. The Streaming API works for registered users only, but we would be more than willing to take care of that if someone decided to write the needed connector
More...
 
Cannot comment on the Metlab side of things, but we really tried to make our API very simple. If you find an experienced programmer to deal with it and they need some assistance, please have them contact us.
 
You must log in or register to reply here.
Back
Top