No benchmark results are available. Run
Rscript bench/run.R --suite=full from the package root to
create them.
Quality
Quality is measured on a random subsample of at most 5,000 points. Trustworthiness measures whether neighbours in the embedding are also neighbours in the input space; kNN accuracy is the accuracy of a 5-nearest neighbour vote on the known labels in the embedding. Equal values mean the implementations find embeddings of equal quality, as expected from faithful ports of the same algorithm.