[R-pkgs] randomForestSRC 2.0.0
Udaya B. Kogalur
kogalurshear at gmail.com
Sun Dec 13 17:50:45 CET 2015
randomForestSRC 2.0.0 is now available on CRAN. The package is able
to handle multivariate regression (data sets in which the response is
one or more continuous variables), multivariate classification (data
sets in which the response is one or more factor variables),
multivariate mixed (data sets in which the response is a combination
of continuous and factor variables), survival, competing risk, and
unsupervised forests (data sets in which no response is chosen).
The memory and CPU profiles for the package have improved. The
package is OpenMP compliant on all popular platforms. Instructions
for enabling this functionality is available in Rd files, and
pre-compiled OpenMP binaries are also available at:
Linear gains in computation times have been observed on multi-core
desktops and cluster environments using OpenMP.
Finally, the ability to implement user defined split rules is now
available. Instructions and examples for all families are provided in
the package's src directory in splitCustom.c.
Udaya B. Kogalur, Ph.D.
Contract Staff, Dept. of Quantitative Health Sciences, Cleveland
ubkogalur at gmail.com
More information about the R-packages