README.md

In mschubert/bhtsneR: Barnes-Hut t-SNE wrapper

bhtsneR

A minimal, CRAN-compliant wrapper for the Barnes-Hut implementation of T-distributed Stochastic Neighbor Embedding (t-SNE).

Installation

You will need the following software:

• cblas or openblas
• R packages: Rcpp, devtools

Run this in R to install the current GitHub version:

devtools::install_github("mschubert/bhtsneR")

Running

Here is an example on how to run it on the iris data set:

library(bhtsneR)

# perform Barnes-Hut t-SNE
result = bhtsne(iris[1:4])
plot(result, col=iris[[5]])

# perform exact calculations
# this is equivalent to bhtsne(..., theta=0)
result = tsne(iris[1:4])
plot(result, col=iris[[5]])

More information

Barnes-Hut-SNE. (2013). L.J.P. van der Maaten. arXiv:1301.3342.

mschubert/bhtsneR documentation built on May 23, 2019, 7:50 a.m.