rangeFind-methods: Find all neighbors in range
In BiocNeighbors: Nearest Neighbor Detection for Bioconductor Packages

Description Usage Arguments Details Value Author(s) See Also Examples

Find all neighboring data points within a certain distance of each point.

rangeFindKmknn(X, threshold, get.index=TRUE, get.distance=TRUE,
    BPPARAM=SerialParam(), precomputed=NULL, subset=NULL,
    raw.index=FALSE, ...)

rangeFindVptree(X, threshold, get.index=TRUE, get.distance=TRUE,
    BPPARAM=SerialParam(), precomputed=NULL, subset=NULL,
    raw.index=FALSE, ...)

rangeFindExhaustive(X, threshold, get.index=TRUE, get.distance=TRUE,
    BPPARAM=SerialParam(), precomputed=NULL, subset=NULL,
    raw.index=FALSE, ...)

`X`	A numeric matrix where rows correspond to data points and columns correspond to variables (i.e., dimensions).
`threshold`	A positive numeric scalar specifying the maximum distance at which a point is considered a neighbor. Alternatively, a vector containing a different distance threshold for each point.
`get.index`	A logical scalar indicating whether the indices of the neighbors should be recorded.
`get.distance`	A logical scalar indicating whether distances to the neighbors should be recorded.
`BPPARAM`	A BiocParallelParam object indicating how the search should be parallelized.
`precomputed`	A BiocNeighborIndex object of the appropriate class, generated from `X`. For `rangeFindExhaustive`, this should be a ExhaustiveIndex from `rangeFindExhaustive`. For `rangeFindKmknn`, this should be a KmknnIndex from `rangeFindKmknn`. For `rangeFindVptree`, this should be a VptreeIndex from `rangeFindVptree`.
`subset`	A vector indicating the rows of `X` for which the neighbors should be identified.
`raw.index`	A logial scalar indicating whether raw column indices should be returned, see `?"BiocNeighbors-raw-index"`.
`...`	Further arguments to pass to the respective `build*` function for each algorithm. This includes `distance`, a string specifying whether `"Euclidean"` or `"Manhattan"` distances are to be used.

This function identifies all points in X that within threshold of each point in X. For Euclidean distances, this is equivalent to identifying all points in a hypersphere centered around the point of interest. The exact implementation can either use the KMKNNN approach or a VP tree.

By default, a search is performed for each data point in X, but it can be limited to a specified subset of points with subset. This yields the same result as (but is more efficient than) subsetting the output matrices after running findNeighbors with subset=NULL.

If threshold is a vector, each entry is assumed to specify a (possibly different) threshold for each point in X. If subset is also specified, each entry is assumed to specify a threshold for each point in subset. An error will be raised if threshold is a vector of incorrect length.

Turning off get.index or get.distance will provide a slight speed boost and reduce memory usage when these returned values are not of interest. If both get.index=FALSE and get.distance=FALSE, an integer vector containing the number of neighbors to each point is returned instead, which is more memory efficient when the identities of/distances to the neighbors are not required.

Using BPPARAM will parallelize the search across points, which usually provides a linear increase in speed.

If multiple queries are to be performed to the same X, it may be beneficial to build the index from X (e.g., with buildKmknn). The resulting BiocNeighborIndex object can be supplied as precomputed to multiple function calls, avoiding the need to repeat index construction in each call. Note that when precomputed is supplied, the value of X is ignored.

A list is returned containing:

index, if get.index=TRUE. This is a list of integer vectors where each entry corresponds to a point (denoted here as i) in X. The vector for i contains the set of row indices of all points in X that lie within threshold of point i. Points in each vector are not ordered, and i will always be included in its own set.
distance, if get.distance=TRUE. This is a list of numeric vectors where each entry corresponds to a point (as above) and contains the distances of the neighbors from i. Elements of each vector in distance match to elements of the corresponding vector in index.

If get.index=FALSE and get.distance=FALSE, an integer vector is returned instead containing the number of neighbors to i.

If subset is not NULL, each entry of the above lists corresponds to a point in the subset, in the same order as supplied in subset.

See ?"BiocNeighbors-raw-index" for an explanation of the output when raw.index=TRUE.

Aaron Lun

buildExhaustive, buildKmknn or buildVptree to build an index ahead of time.

See ?"BiocNeighbors-algorithms" for an overview of the available algorithms.

Y <- matrix(runif(100000), ncol=20)
out <- rangeFindKmknn(Y, threshold=3)
out2 <- rangeFindVptree(Y, threshold=3)
out3 <- rangeFindExhaustive(Y, threshold=3)