An Approach to Parallelizing Isotonic Regression
Anthony J. Kearsley, Richard A. Tapia, Michael W. Trosset · 1996
Isotonic regression is the problem of fitting data to order constraints. We demonstrate that the isotonic regression of a finite set of numbers Y can be obtained by decomposing Y into subsets, performing parallel isotonic regressions on each subset, then performing a trivial isotonic regression on the resulting combined set. Numerical experiments confirm the efficacy of this approach. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.