MDA: Maximum Dissimilarilty Algorithm

In OakleyJ/MUCM: Gaussian process emulator methods based on the MUCM toolkit

Description

Selects a subset of events by maximizing the dissimilarity between selected samples. Supports biased selection via event weighting. Does not require starting events to have been selected from same population or with same algorithm.

Usage

MDA(data.selected, data.new, n, weight.selected = NULL, weight.new = NULL,
  normalise = TRUE, index = FALSE)

Arguments

data.selected	a matrix (or data.frame) of events already selected.
data.new	a matrix (or data.frame) from which to select new events from.
n	the number of additional events to be selected from data.new.
weight.selected	optional argument (vector) used to weight the events already selected (paired with events in data.selected, and should have length = nrow(data.selected)).
weight.new	optional argument (vector) used to weight the events to be selected (paired with events in data.new, and should have length = nrow(data.new)).
normalise	whether to first normalise the data so each column is scaled between 0 and 1. Defaults to TRUE but use FALSE if already normalised.
index	a logical argument. If FALSE (default), the function returns a matrix (or data.frame) of selected events. If TRUE it returns indices of the selected events in data.new.

Value

If index = FALSE the function returns a matrix (or data.frame) of selected events. Otherwise it returns the indices of the selected events in data.new named with row names if available.

Author(s)

Sajni Malde

OakleyJ/MUCM documentation built on May 7, 2019, 9:01 p.m.