despeckle {beadplexr} | R Documentation |
Remove lonely, noisy data points in a 2D scatter matrix
despeckle(df, .parameters, .bins = 256L, .neighbours = 4L, .data = NULL, ...)
df |
A tidy data.frame |
.parameters |
A character of the length of two giving the parameters to despeckle. |
.bins |
A numeric giving the resolution of the raster matrix. |
.neighbours |
A numeric giving the minimum number of neighbors. Points with fewer neighbors are removed. |
.data |
Deprecated. Use |
... |
Additional parameters passed to |
The values of the two parameters are binned into the given number of bins.
They are then cast into a 2D matrix, with the bins of the first of the
parameters ending up as rows, the bins of the second parameter as
columns, and combinations are marked by 1
.
This matrix is turned into a RasterLayer
by raster::raster()
and the
number of neighbors are calculated by raster::clump()
.
The rows of the df
where lonely points are found in .parameters
are removed.
A data.frame
with noisy points removed.
This function requires that the igraph
package is available.
library(beadplexr) library(magrittr) library(ggplot2) data("lplex") lplex[[1]] %>% ggplot() + aes(x = `FL6-H`, y = `FL2-H`) + geom_point() lplex[[1]] %>% despeckle(.parameters = c("FL6-H", "FL2-H"), .neighbours = 8) %>% ggplot() + aes(x = `FL6-H`, y = `FL2-H`) + geom_point() lplex[[1]] %>% despeckle(.parameters = c("FL6-H", "FL2-H"), .bin = 128, direction = 4) %>% ggplot() + aes(x = `FL6-H`, y = `FL2-H`) + geom_point()