Optimal sub-Gaussian variance proxy for 3-mass distributions
By: Soufiane Atouani, Olivier Marchal, Julyan Arbel
Potential Business Impact:
Makes math tools better for guessing random events.
We investigate the problem of characterizing the optimal variance proxy for sub-Gaussian random variables,whose moment-generating function exhibits bounded growth at infinity. We apply a general characterization method to discrete random variables with equally spaced atoms. We thoroughly study 3-mass distributions, thereby generalizing the well-studied Bernoulli case. We also prove that the discrete uniform distribution over $N$ points is strictly sub-Gaussian. Finally, we provide an open-source Python package that combines analytical and numerical approaches to compute optimal sub-Gaussian variance proxies across a wide range of distributions.
Similar Papers
Sub-Poisson distributions: Concentration inequalities, optimal variance proxies, and closure properties
Probability
Makes math tools work better for predicting random events.
A variational approach to dimension-free self-normalized concentration
Probability
Makes math work better for smart computer learning.
Sharp empirical Bernstein bounds for the variance of bounded random variables
Statistics Theory
Makes computer guesses about data more reliable.