Embed a graph into a UMAP, Uniform Manifold Approximation and Projection for Dimension Reduction, <https://github.com/lmcinnes/umap>, <doi:10.21105/joss.00861>
embedGraphUmap(
graph,
min.prob = 0.001,
min.visited.verts = 1000,
n.cores = 1,
max.hitting.nn.num = 0,
max.commute.nn.num = 0,
min.prob.lower = 1e-07,
n.neighbors = 40,
n.epochs = 1000,
spread = 15,
min.dist = 0.001,
return.all = FALSE,
n.sgd.cores = n.cores,
verbose = TRUE,
...
)
input igraph object
numeric Minimum probability for proximity when calculating hitting time per neighbors (default=1e-3)
numeric Minimum number of vertices visted when calculating hitting time per neighbors (default=1000)
numeric Number of cores to use (default=1)
numeric Maximum adjacencies for calculating hitting time per neighbor, hitting_time_per_neighbors() (default=0)
numeric Maximum adjacencies for calculating commute time per neighbor, commute_time_per_node() (default=0)
numeric Probability threshold to continue iteration in depth first search hitting time, dfs_hitting_time() (default=1e-7)
numeric Number of neighbors (default=40)
numeric Number of epochs to use during the optimization of the embedded coordinates (default=1000). See 'n_epochs' in uwot::umap()
numeric The effective scale of embedded points (numeric default=15). See 'spread' in uwot::umap()
numeric The effective minimum distance between embedded points (default=0.001). See 'min.dist' in uwot::umap()
boolean If TRUE, return list(adj.info=adj.info, commute.times=commute.times, umap=umap). Otherwise, just return UMAP(default=FALSE)
numeric Number of cores to use during stochastic gradient descent. If set to > 1, then results will not be reproducible, even if 'set.seed' is called with a fixed seed before running (default=n_threads) See 'n_sgd_threads' in uwot::umap()
boolean Verbose output (default=TRUE)
Additional arguments passed to embedKnnGraph()
resulting UMAP embedding