x2p returns a pair-wise conditional probability matrix given an input
matrix X.
x2p(X, perplexity = 30, tol = 1e-05)A data matrix with \(N\) rows.
The target perplexity. Values between 5 and 50 are generally considered appropriate. Loosely translates into the expected number of neighbors per point.
A small positive value.
An N x N matrix of pair-wise probabilities.
This function is an almost direct port of the original Python implementation
by van der Maaten and Hinton (2008). It uses a binary search to estimate
probability values for all pairwise-elements of X. The conditional
Gaussian distributions should all be of equal perplexity.
L.J.P. van der Maaten and G.E. Hinton. ``Visualizing High-Dimensional Data Using t-SNE.'' Journal of Machine Learning Research 9(Nov):2579-2605, 2008. PDF.