Yin-Ting Liao: Sharp large deviation estimates for random projections of $\ell^n_p$ balls.

Abstract

Random projections of high-­dimensional probability measures have gained much attention in asymptotic convex geometry and high­ dimensional statistics. Accurate estimate of the tail probabilities is of importance in applications. The large deviation principles for such projections has only been recently studied. We further refine the estimates to obtain the prefactor apart from the exponential decay in the spirit of Bahadur and Ranga-Rao. In this talk, I will describe our results on sharp large deviation estimates for random projections of \(\ell^n_p\) balls. Both theoretical and computational results would be provided.

This talk is based on several joint works with Kavita Ramanan.