Matlab in Finance

Quote from linuxtrader:
...incurr the penalty involved in a matlab component...

Our last native matlab and mathematica applications are scheduled to be phased out by the end of the year.
More...
What penalty?

C/C++ is easier to prototype with? Perhaps you can say this if you have an extensive testing and trading infrastructure already in place. Otherwise anyone doing hard core math operations will benefit greatly from Matlab and Mathematica. This is especially true when prototyping and testing ideas, and needing to visualize information effectively.
 
Quote from prophet:

What penalty?

C/C++ is easier to prototype with? Perhaps you can say this if you have an extensive testing and trading infrastructure already in place. Otherwise anyone doing hard core math operations will benefit greatly from Matlab and Mathematica. This is especially true when prototyping and testing ideas, and needing to visualize information effectively.
More...

Uhh ... I said prototype (new ideas) in mathematica or matlab (less preferred). ... And yes, matlab incurrs a performance penalty for most operations that we are interested in ..... We rarely use it for real work.

Large simulations (my idea of hard core math and computer systems engineering) are not done in either environment ......
 
I must have misinterpreted part of what you wrote. Sorry.

Regarding the performance penalty with Matlab… how are you using Matlab? Real time processing of market data? I agree there is too much overhead for RT work. However, if you are using Matlab for model development or backtesting, you can always use vector operations and MEX functions to increase performance to nearly the same as C/C++, for most operations.
 
I want to study groups of data, that is; create distributions (either on a fixxed amount of time or volume, or based on an event creating a start point) and then save the descriptive statistics of each distribution for data mining.

Instead of looking at individual bars (5min, 30min, daily) with high, low and close; I want to look at successive distributions as defined by mean, std dev, kurtosis, skew, etc. I don't necessarily need a graphic representation of this, but it would be nice.

Any ideas, or if I"m not being clear enough, I'll try and be more specific?
 
So instead of just having OHLC + volume per bar, you'd like to see some other information, per bar, like recent volatility, VWAP, or S/R estimates? You can do this per-tick as well, though per-bar is easier to start with. If this is your goal you can certainly do it with Matlab and Mathematica, and probably with a lot other software too.
 
prophet, yep, like what you said.

The biggest challenge is I can't program (at least not yet). I don't have a concept of how to group the data. I think the code guys call it a data structure.

Should I think about creating an array for each bar? The array could hold all the stat's.

Thanks much for your reply.
 
Hi,

I am not going to advertise here but why don't you take a look at QuantStudio, www.smartquant.com. We have designed QuantStudio to "substitute" MatLab in financial data analysis. In fact QuantStudio has its roots in the ROOT package : root.cern.ch designed for the analysis of high energy and nuclear data... The ROOT package is widely used by scientific community and is an open source c++ package...

Regards,
Anton
 
You must log in or register to reply here.
Back
Top