Paperlane

no. 16 · Tuesday 17 November 2026

Stochastic Multivariate Taylor Measure Functions

Athanasios Christou Micheas

math.PR · 63 pp

Abstract

We introduce stochastic multivariate Taylor measure functions (SMTMFs), which extend the multivariate Taylor measure function framework through random Taylor coefficients and a possibly random Taylor parameter. We establish their measurability properties and derive sufficient conditions for the associated processes to be martingales, including criteria based on conditional covariance and continuous semimartingale decompositions. We construct standard Brownian motion as an almost surely uniformly convergent limit of finite SMTMFs using the Levy-Ciesielski representation. We also show that the completion of the finite stochastic MTMF space under the expected uniform norm contains Brownian motion, Brownian bridges, fractional Brownian motion with Hurst parameter H, and Ornstein-Uhlenbeck processes with integrable initial conditions. Our main result characterizes this completion as the Bochner space, under an admissibility condition on jointly measurable stochastic Taylor coefficients. The proof uses dyadic piecewise linear interpolation to approximate integrable continuous stochastic processes in the expected uniform norm. Thus, the completed SMTMF space encompasses the full class of integrable continuous random functions on a compact time interval.

comments
63 pages; Keywords: Bochner space; Brownian motion; fractional Brownian motion; Martingales; Semimartingales; Stochastic multivariate Taylor measure functions
fields
Probability, Functional Analysis, Statistics Theory
pdf ↗ page arXiv:2611.20027 ↗