dataPrep: Neighborhood Data Preparation

Description

Prepares a neighborhood data from a given image, using the left-to-right then top-to-bottom raster scan order (see Bui and Apley 2018a).

Usage

dataPrep(img, nb, vars = NULL, subsample = 1)

Arguments

img

the given image in the matrix format.

the size of the neighborhood. It must be a 1-length or 3-length vector of positive integer(s). If the former, it is the same with a 3-length vector with the same elements.

vars

names of variables to be selected in the neighborhood data.

subsample

the portion of data rows be returned. It takes values in (0, 1]. If subsample = 1, all data rows will be returned, and if subsample = .5, only roughly a half will be returned.

Value

A dataframe with column names "V1", "V2", "V3", ... The first column "V1" contains the response pixel, whereas the other columns contain pixels in the neighborhood (with size nb) of the response pixel.

References

Bui, A.T. and Apley., D.W. (2018a) "A Monitoring and Diagnostic Approach for Stochastic Textured Surfaces", Technometrics, 60, 1-13.

Examples

Run this code

# NOT RUN {
## construct a neighborhood data for an unrealistically small mock image (7x9 pixels).
mock.img <- matrix(sample(0:255, 63, replace = TRUE), 7, 9)
mock.img
dataPrep(img = mock.img, nb = 2) # the same with nb = c(2, 2, 2)

## select only columns "V2", "V5", and "V13" in the output
dataPrep(img = mock.img, nb = 2, vars = c("V2", "V5", "V13"))

## return only a half number of rows
dataPrep(img = mock.img, nb = 2, subsample = .5)
# }