T-distributed stochastic neighbor embedding: Difference between revisions

Revision as of 05:50, 27 February 2023

Revision as of 05:53, 27 February 2023

Line 27:

This is motivated because <math>p_{i}</math> and <math>p_{j}</math> from the N samples are estimated as 1/N, so the conditional probability can be written as <math>p_{ji\mid ij} = Np_{ij}</math> and <math>p_{j\mid i} = Np_{ji}</math> . Since <math> p_{jiij} = p_{ji}</math>, you can obtain previous formula.

Also note that <math>p_{ii} = 0 </math> and <math>\sum_{i, j} p_{ij} = 1</math>.