Releases: jlmelville/uwot
Releases · jlmelville/uwot
v0.0.0.9004
New features
- For supervised UMAP with numeric
y
, you may pass nearest neighbor data
directly, in the same format as that supported byX
-related nearest neighbor
data. This may be useful if you don't want to use Euclidean distances for
they
data, or if you have missing data (and have a way to assign nearest neighbors
for those cases, obviously). See the Nearest Neighbor Data Format section for details.
v0.0.0.9003
New features
- New parameter
ret_nn
: whenTRUE
returns nearest neighbor matrices
as ann
list: indices in itemidx
and distances in itemdist
. Embedded
coordinates are inembedding
. Bothret_nn
andret_model
can beTRUE
,
and should not cause any compatibility issues with supervised embeddings. nn_method
can now take precomputed nearest neighbor data. Must be a list of
two matrices:idx
, containing integer indexes, anddist
containing
distances. By no coincidence, this is the format return byret_nn
.
Bug fixes and minor improvements
- Embedding to
n_components = 1
was broken (#6) - User-supplied matrices to
init
parameter were being modified, in defiance of basic R pass-by-copy semantics.
v0.0.0.9002
Better late then never, I will start creating releases, this one back-dated to the middle of August.
metric = "cosine"
is working again forn_threads
greater than0
(#5)