Doubly Outlier-Robust Online Infinite Hidden Markov Model

Horace Yiu, Leandro Sánchez-Betancourt, Álvaro Cartea, Gerardo Duran-Martin · 2026

Read Doubly Outlier-Robust Online Infinite Hidden Markov Model by Horace Yiu, Leandro Sánchez-Betancourt, Álvaro Cartea, Gerardo Duran-Martin (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-14322.txt.

Introduction

Yiu, Sánchez-Betancourt, Cartea, and Duran-Martin derive a robust update for the online infinite hidden Markov model for streaming data with outliers and a misspecified model. Robustness is defined through the posterior influence function, with conditions for bounded influence, though robustness necessarily slows regime switching. Their batched robust variant balances adaptation speed against robustness with two tunable parameters. On limit order book data, hourly power demand, and a synthetic high-dimensional system, one-step-ahead forecast error drops by up to 67 percent against rival online Bayesian methods.

Catalog record

CREATORHorace Yiu, Leandro Sánchez-Betancourt, Álvaro Cartea, Gerardo Duran-Martin
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE180344
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.14322

Browse the full library · About the Tower of Babel