Matlab for Beginners

[kamikaze_trader]

Hi I'm new here in this forum and in trading world (also very happy to be here). i'm looking for books and websites about "trading with matlab" for beginners. I feel like I'm losing too much time to learn very basic stuff (like pulling data from yahoo and formatting it for instance). Especially, historical price data gets messed up when I use datafeed toolbox. Please help, my head's gonna explode. Thanks

[henry145]

Video Tutorials



Interactive brokers via Matlab



hope that helps

[rafael_nicolas]

Check the blog of Dr. Chan

http://epchan.blogspot.com/2009/05/matlab-as-automated-execution-system.html



Also you can find some cool datafeed toolbox: www.rnfc.org/ivey/



nico

[chiral3]

This is always a good one



www.maths.strath.ac.uk/~aas96106/39326.pdf

[kamikaze_trader]

Awesome!! Thanks all of you for the feedback! Smiley



The codes were really helpful, but looks like i need to work more on the fundamentals for now. Adios!

[lehalle]

Attilio's book is really a good way to begin:

Risk and Asset allocation

[jungle]

Trading with MATLAB (think this is the website of a phorum member?)



There used to be a site matlabtrader.com, but the link appears to be broken.



For books, try Numerical Methods In Finance And Economics: A MATLAB-Based Introduction.

[Breno Neri]

Seriously, forget about Matlab. Go with R (r-project.org) instead. It's open source and it's so much powerful than Matlab. There are hundreds of packages available for download.



That's what statisticians have been using. You can do so much stuff in one single line. Jarque-Bera test for normality, Kolmogorov-Smirnoff test for adherence, Johansen's test for cointegration, Dickey-Fuller test for unit root, wilcoxon's test, it's all there.



There are so many different regressions you can perform in one line, from the vanilla least squares regression to ordered probit to quantile regression, etc.



For any probability distribution you can think of (normal, t, gamma, beta, von-mises, cauchy, exponential, alpha stable, and other dozens), R has four commands: random number generator, cdf, pdf (or pmf, if it's a discrete distribution) and quantile function (the inverse of the cdf).



It can access sql servers and kdb servers (kdb is much faster than relational data bases like sql for tick-by-tick data. That's what all the big high frequency shops use, but this is a topic for another post).



In summa, R is so much more complete than Matlab. It's definitely worth taking a look at.

[FDAXHunter]

That is complete drivel. R is a statistical tool, whereas Matlab is an engineering/computer science tool. R is the open source version of S and no serious person in finance (let alone trading) would ever consider using S.



Matlab can also access pretty much any kind of database (including obviously SQL Server and kdb(+)). It can also interact with many real-time data feeds and protocols.



Whom in the actual world of finance do you know that uses Kolmogorov-Smirnoff or even Jarque-Bera? These are academic tests and while perhaps of great interest to the statistical community, the real world doesn't care. Do you work in academia or as you so eloquently put it: "a high-frequency shop"?



Also, I note that R is always slower than Matlab, no matter what you do.

[Maggette]

Even though the "R vs. Matlab" debate is getting lame, I agree with FDAX. Stating that R is more powerful is complete nonsense. Just not true. R is free, has a very well working community, has a large bunch of nice and good working packages that are easy to install and to use, and is a very, very, very nice tool .....but in terms of computational power,R is not even close to MATLAB

...cheers