AI & ML
Parameter estimation for discretely observed rough fractional interacting particle systems
arXiv stat.TH0 views
arXiv:2611.24287v1 Announce Type: new Abstract: We study parametric inference for interacting particle systems driven by independent fractional Brownian motions, based on discrete observations of all particles over a fixed time horizon. Our goal is to estimate an unknown drift parameter in the joint limit as the number of particles increases and the observation mesh vanishes. For $H>\frac 12$, we construct contrast based estimators by extending the discrete time least squares approach of Neuenkirch and Tindel to the interacting setting. For the rough regime $H\in(\frac 14,\frac 12)$, we introduce a modified contrast and use the stochastic sewing lemma to handle the first order approximations available from discrete observations. Our analysis combines the propagation of chaos for the associated fractional McKean Vlasov equation with estimates controlling the temporal discretization error. When the Hurst parameter and the diffusion coefficient are known, we establish strong consistency and asymptotic normality of the proposed estimators. The asymptotic results reveal an interplay between the number of particles and the temporal discretization: the convergence rate with respect to the observation mesh changes at $H=\frac 34$, while the asymptotic normality holds in both regimes considered. We further develop a joint estimation procedure for the drift parameter, the Hurst parameter, and the diffusion coefficient, extending quadratic-variation methods for fractional Brownian motion to the interacting particle framework. Numerical experiments illustrate the performance of the proposed estimators.
Source:https://arxiv.org/abs/2611.24287PreviousBerry Esseen bounds for fractional Vasicek type processes: continuous and high-frequency regimesNextWorst-case bias of the independence assumption for functions of two ordinal variables
Related posts
Multimodal Masked Diffusion Autoencoder for Modeling Impaired Driving
0arXiv cs.RO
CoNav-UAV: Cooperative Dual-Altitude Aerial Navigation via Stackelberg Learning
0arXiv cs.RO
Meta-RL with Bayesian Linear Task Models
0arXiv cs.RO
RFW-S: Everyday Deformable Objects with Independent Simulation Verification
0arXiv cs.RO
Heave-Ho: Decentralized Long-Horizon Multi-Humanoid Collaboration
0arXiv cs.RO
CriticHack: Evaluating Visual Rewards Under Robot Policy Optimization
0arXiv cs.RO