peperr 1.4

peperr 1.3

Suggestion of package CoxBoost and corresponding functions removed, as this package was archived.

Suggestion of package randomSurvivalForest resolved, as this package was archived.

bug concerning weighting in censored competing risks situation in function pmpec() fixed

New peperr() arguments ‘sr’, ‘sr.name’ and ‘sr.restore’ to allow for restoring previously saved results
Incorporation of interface to package randomSurvivalForest: Functions fit.rsf_mtry, complexity.ipec.rsf_mtry allow for selection of number of candidate values at each node by IPEC

New function ipec(), to caclulate the integrated prediction error curve for survival data

Default value of peperr argument RNG changed to ‘RNGstream’
New peperr() argument lb, to choose if computation on slaves should be executed load balanced

Independent parallel random number streams integrated -> new peperr() argument RNG
peperr_1.1.tar.gz contained peperr_1.0-3.tar.gz by mistake
plot.peperr() now works for non-integer complexity values as well

New arguments of function peperr() are ‘clustertype’, to specify the type of cluster to be used (‘SOCK’ for socket, ‘MPI’, ‘PVM’ or ‘NWS’) and ‘clusterhosts’, that specifies the host list for socket clusters, if ‘clustertype=SOCK’ or ‘clustertype=NWS’.
Function extract.fun() returns additional list element ‘variables’, containing global variables required for computation on cluster nodes.
Default of argument ‘parallel’ in function peperr() changed from ‘FALSE’ to ‘NULL’. This fixed bug that switching from parallel to sequential execution was not possible, but is not affecting any other cases.

New argument ‘seed’ in function peperr(), to allow reproducibility in parallel execution