Non-Homogeneous Markov-Switching Generalized Additive Models for Location, Scale, and Shape

Katharina Ammann 1 , Timo Adam 1 , Jan-Ole Koslik 1 1 Bielefeld University, Germany · 2026

Read Non-Homogeneous Markov-Switching Generalized Additive Models for Location, Scale, and Shape by Katharina Ammann 1 , Timo Adam 1 , Jan-Ole Koslik 1 1 Bielefeld University, Germany (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-2601-03760.txt.

Description

Katharina Ammann, Timo Adam, and Jan-Ole Koslik extend a class of statistical models used for time series that switch between hidden regimes. Their models combine distributional regression, which can model a distribution's location, scale, and shape, with hidden Markov switching between states. Until now, the odds of switching states were usually held constant over time, so regime changes could not respond to outside conditions.

The extension lets covariates drive the transition probabilities themselves, modeled as smooth functions. Estimation uses penalized likelihood with automatic smoothness selection to guard against overfitting. Simulations plus applications to daily Lufthansa stock prices and Spanish energy prices test the method. Markets and energy systems both switch regimes in response to conditions, so letting the switching itself react to covariates is a natural, useful step.

Catalog record

CREATORKatharina Ammann 1 , Timo Adam 1 , Jan-Ole Koslik 1 1 Bielefeld University, Germany
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE42 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-03760

Browse the full library · About the Tower of Babel