ch.distOverlapMaxAveCombXY: Bootstrap the overlap of two distributions Function with a...

In ccpluncw/ccpl_R_chValues:

Bootstrap the overlap of two distributions Function with a flexible function

Description

This function does a bootstrap to calculate the percentage of samples that Distribution X is greater than Distribution Y. When Distribution X (or Y) are the combination of multiple items, then it uses combFun (a function) to combine the values of those items into a single distribution.

Usage

ch.distOverlapMaxAveCombXY(
  xDist,
  yDist,
  numRuns,
  probMaxX = 0.5,
  probMaxY = 0.5
)

Arguments

xDist	a list containing the values for Distribution X.
yDist	a list containing the values for Distribution Y.
numRuns	the number of runs in the bootstrap. DEFAULT = 100.
probMaxX	A probability that specifies the probability of samples that should be devoted to the xDist list column with the max(median()). When probMax = 0, the column with the max(median()) has no influence on either the average or above and beyond the average. When probMax = 0.5, the column with the max(median()) is 50 percent of the output vector. This is simply collapsing into a vector (as.vector) when there are only 2 vectors in the list. But it will have an outsized influence if there are more than two vector in the list. When probMax = 1, the column with the max(median()) is 100 percent of the output vector. DEFAULT = 0.5 (equal influence for two vectors)(probMax = .8 will produce a distribution in which .4 of the total are all the items (including the max) and .6 of the total are MAX items - when there are only two vectors in the list). Use ch.getProbMax() to calculate the appropriate probMax.
probMaxY	The same as probMaxX, but for the yDist list.

Value

the average percentage of samples that xDist was greater than yDist, the sd of the average percentage, Overlap = a transformation of average percentage so 1=complete overlap and 0 = no overlap, direction = 1 if xDist > yDist; else -1

Examples

ch.distOverlapMaxAveCombXY (xValue, yValue,100, combFun = ch.maxAveComb, probMax = 0.667)

ccpluncw/ccpl_R_chValues documentation built on Feb. 12, 2024, 4:21 a.m.