Strongly convex stochastic online optimization on a unit simplex with application to the mixing least square regression

TL;DR

This paper introduces a novel stochastic online mirror descent method for strongly convex optimization on a unit simplex, with applications to mixing least squares regression, enhancing estimation accuracy in non-Euclidean settings.

Contribution

It presents a new approach for mixing least squares regression using stochastic online mirror descent tailored for non-Euclidean spaces.

Findings

Effective estimation on the unit simplex

Improved convergence in non-Euclidean settings

Applicability to mixing least squares regression

Abstract

In this paper we propose a new approach to obtain mixing least square regression estimate by means of stochastic online mirror descent in non-euclidian set-up.