[R-SIG-Finance] quantmod/Rmetrics-timeSeries and RQuantLib/Rmetrics-fOptions ?

[Brian G. Peterson]

On Thu, 2011-05-05 at 02:49 -0700, msalese wrote:
> Hi guys,
> I'm a no-pro option trader and I'm learning to use R + packages that can
> help me to do better analysis and quit spreadsheets.
> I bought the book "R in a Nutshell" and booked the on-line classes from
> statistics.com on R language  (introduction classes will start on 20 of
> May).

Paul Teetor's book is excellent as well: 'R Cookbook'

I can't speak to the quality or lack thereof of any commercial R
training.  When evaluating it, I would suggest that you evaluate the
published contributions of the author or instructor to R.

> There are some doubts on which I need some help, I'm going to spent a lot of
> time on learning steeps and R has a lot of packages for time series analysis
> and options, the point is on which packages it's better to stay focused ?
> a. quantmod or timeSeries/Rmetrics ?

These are very different things.

Rmetrics has a huge number of functions, many of them useful.  For
those, you will need timeSeries data.

xts is the fastest, highest capacity time series class in R. Many of us
use xts with huge amounts of tick data (tens of billions of
observations).  This is observable, verifiable, fact. Period. xts is
also most easily convertible to and from other time series classes and
other R object types so that you can most easily use other packages in
R.

quantmod is mostly used for the excellent charting chartSeries and
chart_Series addTA/add_TA.  getSymbols is also extremely useful, even
with custom proprietary data sources.
  
> b. RQuantLib or fOptions/Rmetrics ?

fOptions is likely easier to learn, in no small part because of the
ebooks you mention below.  As with all of the Rmetrics suite, it is
primarily aimed at *teaching* and *research*, not necessarily
application to real markets.

quantlib and RQuantlib are widely reviewed and used by many
institutional investors, so you can believe that the calculations in the
quantlib library have been extensively checked for correctness.

> c. Do you recommend eBooks from Rmetrics ?

The portfolio optimization book, the only one I have a copy of, is good
as far as it goes, but lacks more advanced topics.

HTH,

 - Brian

-- 
Brian G. Peterson
http://braverock.com/brian/
Ph: 773-459-4973
IM: bgpbraverock