neighbors_by_dist: Determine and store neighbors by Euclidean Distance...

In jmleach-bst/sim2Dpredictr: Simulate Outcomes Using Spatially Dependent Design Matrices

Determine and store neighbors by Euclidean Distance Constraints

Description

Determine and store neighbors by Euclidean Distance Constraints

Usage

neighbors_by_dist(x, y, coords, im.res, r, print.ring = FALSE)

Arguments

x, y	are the row and column coordinates, respectively.
coords	A dataframe containing indices and coordinates for the image.
im.res	A vector containing the number of rows and columns, respectively.
r	A scalar value determining the radius within which other locations are neighbors to the current location (x, y).
print.ring	When print.ring = TRUE, each iteration is shown, with corresponding information regarding the number of neighbors present in each ring. This argument primarily exists to allow the user to test whether the neighborhood structure specified is as desired.

Value

A tibble whose first column contains x indices, second column contains y indices, and third column denotes the current ring about a location.

Note

This function avoids testing all points for being with a certain distance in order to determine neighbor status of a given point by progressively widening a box around the point. Each iteration widens the box by an extra ring, and we only test points in the new ring. If at the end of testing a ring there are no new neighbors then we stop expanding the box and return the neighbors' coordinates. For computational efficiency, this function assumes that all arguments except the current point's coordinates have been specified.

Examples

## Necessary pre-specified arguments required for the function to work.

## image resoluation + number of spatial predictors
im.res <- c(5, 5)
J <- prod(im.res)

## create predictor indices w/ coordinates
row.id <-rep(1, im.res[2])
for (i in 2:im.res[1]) {
 row.id <- c(row.id, rep(i, im.res[2]))
}
coords <- data.frame(index = 1:J,
                     row.id = row.id,
                     col.id = rep(c(1:im.res[2]), im.res[1]) )

neighbors_by_dist(x = 2, y = 2, im.res = im.res, coords = coords, r = 2)

jmleach-bst/sim2Dpredictr documentation built on March 4, 2024, 5:57 p.m.