NOTE (25/1/2024)
In this fork I am implementing BiocNeighbors
methods for KNN calculation, which run faster than the original covertree
method, and offer a more flexible interface to several other exact (kmknn
, Vptree
) and approximate (hsnw
, already implemented in the original destiny
release, and annoy
) methods. This fork also adds support for sparse matrices for KNN, including rank correlation distance, through matrixStats
.
So far build and tests seem to go well, but some polishing can still be done. I will submit a PR in the near future.
An R package for diffusion maps, with additional features for large-scale and single cell data.