erode.hexbin {hexbin} | R Documentation |
This erosion algorithm removes counts from hexagon cells at a rate proportional to the cells' exposed surface area. When a cell becomes empty, algorithm removes the emptied cell and notes the removal order. Cell removal increases the exposure of any neighboring cells. The last cell removed is a type of bivariate median.
erode.hexbin(bin, cdfcut = 0.5)
bin |
an object of type hexbin. |
cdfcut |
number in (0,1) indicating the confidence level for the limits. |
The algorithm extracts high count cells with containing a given fraction (cdfcut) of the total counts. The algorithm extracts all cells if cdfcut=0. The algorithm performs gray-level erosion on the extracted cells. Each erosion cycle removes counts from cells. The counts removed for each cell are a multiple of the cell's exposed-face count. The algorithm choses the multiple so at least one cell will be empty or have a count deficit on each erosion cycle. The erode vector contain an erosion number for each cell. The value of erode is
6*erosion_cycle_ at_ cell_ removal - cell_deficit_at_removal
Cells with low values are eroded first. The cell with the highest erosion number is a candidate bivariate median. A few ties in erode are common.
A bin object with high count cells and a component called erode that gives the erosion order.
hexbin
, smooth.hexbin
,
hcell2xy
, hcell
, plot.hexbin
,
hboxplot
, hdiffplot
,
hmatplot
, hexagons
, hex.legend
set.seed(153) x <- rnorm(10000) y <- rnorm(10000) bin <- hexbin(x,y) ## bivariate boxplot smbin <- smooth.hexbin(bin) erodebin <- erode.hexbin(smbin,cdfcut=.5) hboxplot(erodebin) # show erosion order oldpar <- par(no.readonly=TRUE) screenpar <- plot.hexbin(bin,style= "lat", minarea=1, maxarea=1, legend=FALSE, border=gray(.7)) par(screenpar) # reset to the plot on the screen hexagons(erodebin,style= "lat", minarea=1, maxarea=1,pen="green") xy <- hcell2xy(erodebin) text(xy, lab = as.character(erodebin$erode),col="white",cex=.65) par(oldpar)