Last chance! 50% off unlimited learning
Sale ends in
This function uses a kd-tree to find the fixed radius nearest neighbors (including distances) fast.
frNN(x, eps, sort = TRUE, search = "kdtree", bucketSize = 10,
splitRule = "suggest", approx = 0)
a data matrix, a dist object or a frNN object.
neighbors radius.
nearest neighbor search strategy (one of "kdtree" or "linear", "dist").
sort the neighbors by distance? This is expensive and can be done later using sort()
.
max size of the kd-tree leafs.
rule to split the kd-tree. One of "STD", "MIDPT", "FAIR", "SL_MIDPT", "SL_FAIR" or "SUGGEST" (SL stands for sliding). "SUGGEST" uses ANNs best guess.
use approximate nearest neighbors. All NN up to a distance of
a factor of 1+approx
eps may be used. Some actual NN may be
omitted leading to spurious clusters and noise points.
However, the algorithm will enjoy a significant speedup.
An object of class frNN (subclass of NN) containing a list with the following components:
a list of integer vectors. Each vector contains the ids of the fixed radius nearest neighbors.
a list with distances (same structure as ids
).
eps used.
For details on the parameters see kNN
.
Note: self-matches are not returned!
To create a frNN object from scratch, you need to supply at least the
elements id
with a list of integer vectors with
the nearest neighbor ids
for each point and eps
(see below).
David M. Mount and Sunil Arya (2010). ANN: A Library for Approximate Nearest Neighbor Searching, http://www.cs.umd.edu/~mount/ANN/.
# NOT RUN {
data(iris)
x <- iris[, -5]
# Find fixed radius nearest neighbors for each point
nn <- frNN(x, eps=.5)
# Number of neighbors
hist(sapply(adjacencylist(nn), length),
xlab = "k", main="Number of Neighbors",
sub = paste("Neighborhood size eps =", nn$eps))
# Explore neighbors of point i = 10
i <- 10
nn$id[[i]]
nn$dist[[i]]
plot(x, col = ifelse(1:nrow(iris) %in% nn$id[[i]], "red", "black"))
# get an adjacency list
head(adjacencylist(nn))
# plot the fixed radius neighbors (and then reduced to a radius of .3)
plot(nn, x)
plot(frNN(nn, .3), x)
## manually create a frNN object for dbscan (dbscan only needs ids and eps)
nn <- list(ids = list(c(2,3), c(1,3), c(1,2,3), c(3,5), c(4,5)), eps = 1)
class(nn) <- c("NN", "frNN")
nn
dbscan(nn, minPts = 2)
# }
Run the code above in your browser using DataLab