For isotonization problems some increase in speed and decrease in complexity can be achieved through the use of the pool monotone groups algorithm of Y.L. Zhang and M.A. Newton (1997). It isotonizes a weighted and ordered set of values. This function is borrowed from package Icens.
Usage
PMGA(est, ww=rep(1, length(est)))
Arguments
est
The vector of values, in the appropriate order.
ww
The weight vector.
Value
An Object containing the following components:
est
The isotonized estimates.
ww
The weight associated with the isotonized estimates.
poolnum
The number of values pooled in the current estimate.
passes
The number of passes which were required to isotonize the list.
Details
To be supplied at some later date.
References
Y. L. Zhang and M. A. Newton (1997), http://www.stat.wisc.edu/~newton/newton.html.