Skip to content

updated Rmetrics packages

1 message · Yohan Chalabi

Original
Yohan Chalabi
# 
Dear list,

We are pleased to announce the update of the Rmetrics packages.

To update a current installation you can use the function
'update.packages()'

Or re-install all Rmetrics packages with the small R script available
at http://www.rmetrics.org:

   source("http://www.rmetrics.org/Rmetrics.R")
   install.Rmetrics()

Most of the work has been focused on 'timeSeries', formerly known as
'fSeries', and 'fPortfolio'.  The other packages have been updated to
be compliant with the new R version 2.9.0. 

* timeSeries

The overall efficiency of 'timeSeries' objects and key functions has
been further improved.

To give you an idea you can try the following example 

   # timeSeries (2100.83)
   library(timeSeries)
   ts <- timeSeries(1:1e6L, 1:1e6L) # one million entries
   ts
   ts[c("1970-01-01 01:00:00", "1970-01-01 12:33:00"), ]
   cbind(ts, ts)
   rbind(ts, ts)
   diff(ts)
   lag(ts)
   ts + ts

Please note that ts[1] returns the first element of the vector ts and
ts[1,] does subset the timeSeries object.

This speed improvement is possible thanks to a careful implementation
of the S4 classes and methods. Note that we are not using additional C
routines. Because of the inner changes, we decided to rename the former
package 'fSeries' to 'timeSeries'. The old package 'fSeries' is still
available on CRAN if you do not want to use the new package.

If you have data sets stored in the former 'timeSeries' definition, you
can use the function '.old2newTimeSeries()' to convert them to the new
format.

* fPortfolio

A lot of work has been done in 'fPortfolio' to improve its overall ease
of use. More specifically, we have added new solvers and new type of
constraints. In-depth information regarding the new features and the
general use of 'fPortfolio' can be found in the forthcoming eBook
"Portfolio Optimization with R/Rmetrics" (expected May 11).

As usual feedback, bug reports and suggestions are always welcome.

Best Regards,

Yohan
For the Rmetrics team


PS: If you are looking for more information, you can meet us in Chicago
at the R/Finance conference or this summer at the Rmetrics Workshop
(more information at http://www.rmetrics.org).