treebin: Tree-based Binning

In rwoldford/treebinr: Data Reduction via Tree-based Binning

Description

treebin bins the provided data using the tree-based binning method, as described in Rahman & Oldford (2018).

Usage

treebin(X, stopCriteria = gapStop, binMeasure = gapMeasure,
  boundaryTest = gapBoundaryTest, selectBin = gapSelect,
  splitBin = gapSplit, makePoint = gapPoints, binInfo = list(binRange =
  matrix(rep(c(-Inf, Inf), ncol(X)), 2, ncol(X))), inputs = list(tau = 1,
  numbins = floor(nrow(X)/2)))

Arguments

X	The point configuration to be binned
stopCriteria	A user supplied function to compute the stopping criteria of the function
binMeasure	A user supplied function to compute the measure associated with each bin
boundaryTest	A user supplied function to test if a given point is contained in a given bin
selectBin	A user supplied function for choosing between bins to be split
splitBin	A user supplied function for splitting a bin
makePoint	A user supplied function to turn the contents of a bin into a single point
binInfo	Additional information to be supplied to the first bin. Default is NULL.
inputs	A list containing additional input parameters required by user supplied functions. Default is NULL.

Value

The return value is an object of class treebinr, which contains the following

points	A matrix containing the reduced point configuration
counts	A vector containing the number of points in each bin
bins	A list containing bin objects, which detail the contents of each bin
tree	An undirected graph object for the binning tree

Examples

set.seed(1324567)
X <- data.frame(x = rnorm(2000), y = rnorm(1000))
out <- treebin(X, inputs = list(tau = 1, numbins = 500))
Xreduced <- as.data.frame(out@points, col.names = c("x", "y"))
Xsampled <- X[sample(1:nrow(X), 500, replace = FALSE),]

savePar <- par(mfrow = c(1,3))
xlim <- extendrange(X$x); ylim <- extendrange(X$y)
plot(X, 
     main = paste0("original data (", nrow(X)," points)"),
     xlim = xlim, ylim = ylim)
plot(Xreduced, 
     main = paste0("reduced data (", nrow(Xreduced)," points)"),
     xlim = xlim, ylim = ylim)
plot(Xsampled, 
     main = paste0("sampled data (", nrow(Xsampled)," points)"),
     xlim = xlim, ylim = ylim)
par(savePar)

rwoldford/treebinr documentation built on May 12, 2019, 4:38 a.m.