# Bandwidth of the gaussian kernels in t-sne

I'm trying to understand t-SNE better and I was hoping someone could elaborate on how the $\sigma _i$'s are chosen.

I was also wondering why they aren't just calculated in the normal way standard deviation is calculated from the data with $x_{i}$ as the expected value for each $\sigma _i$?

• I'm wondering if someone has an answer to this question? How are the standard deviations computed for the Gaussian or t-distribution whose mean is in each point?
– Sos
Aug 3, 2018 at 16:06