A Beta-GAM Hidden Markov Model for Proportion Time Series

Andrea Nigri · 2026

Read A Beta-GAM Hidden Markov Model for Proportion Time Series by Andrea Nigri (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-2605-07300.txt.

Introduction

Andrea Nigri, Han Lin Shang, and Marco Bonetti model proportion time series that switch between hidden regimes. Their hidden Markov model lets a latent state govern spread, while each state's average follows a flexible layered model with spline curves for inputs. The fit alternates expectation and maximization steps under a penalty.

Choosing the model combines information scores with a filter that discards bad solutions, and error bars come from a parametric bootstrap. Test runs recover the switching dynamics well. Applied to Russian mortality ratios across five decades, the model finds two lasting regimes. They fit the country's known mortality shocks, giving the structure a human reading.

Catalog record

CREATORAndrea Nigri
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE59 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.07300

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A Beta-GAM Hidden Markov Model for Proportion Time Series Andrea Nigri Abstract We propose a hidden Markov model for univariate proportion time series taking values in ( 0 , 1 ) (0,1) , where regime switching captures latent structural changes and the emission distribution belongs to the Beta family. In each latent state, the Beta mean is linked to covariates through a generalized additive model (GAM) with spline-based smooth functions, while the Beta precision is state-specific, enabling flexible modeling of both nonlinear covariate effects and regime-dependent variability. Estimation is carried out via a penalized expectation–maximization algorithm, combining smoothing with numerical maximization of the penalized emission likelihood. To select the number of latent states and the smoothing penalty, we implement a grid search guided by standard information criteria (Akaike Information Criterion/Bayesian Information Criterion/Integrated Completed Likelihood) with a diagnostic filter that removes degenerate solutions characterized by explosive precision estimates. Uncertainty is quantified through a parametric bootstrap procedure for transition probabilities and state-dependent parameters. Simulation results demonstrate accurate recovery of transition dynamics, state precisions, and latent-state decoding. A motivating application to Russian age-specific mortality data (1960–2014, ages 0–40) illustrates how the proposed model summarizes smooth age patterns in female-to-total mortality ratios while identifying two persistent latent regimes that admit a substantive demographic interpretation in light of the country’s well-documented mortality shocks that occurred over the second half of the twentieth century. Keywords: Beta regression; Compositional time series; EM algorithm; Generalized additive model; Regime switching model 1 Introduction The presence of proportion data is common in economics, finance, and social statistics, including the income distribution ( 13 ) , financial return distribution ( 14 ) , and the distributions of times when bids are submitted in an online auction ( 12 ) . In demography, one can model a redistribution of normalized life-table death counts over time ( 18 ; 2 , see, e.g.,) . More recently, Beta-distributed emissions have been studied in lockstep with a Hidden Markov model (HMM) and regime-switching additive predictors. On the one hand, 4 proposed a Bayesian Beta-HMM for univariate proportion time series, using conjugate priors and Markov Chain Monte Carlo methods for posterior and predictive inference. Their contribution provides an important starting point for Beta-distributed HMMs, but the emission parameters are constant within each state, and no covariate-dependent structure is introduced. On the other hand, 15 developed Markov-switching generalized additive models (MS-GAMs), showing how state-dependent smooth covariate effects can be estimated by combining HMM with penalized B-splines and maximum penalized likelihood. However, that framework is not specifically designed for Beta-distributed proportion data. Our paper is positioned at the intersection of these two contributions: we combine the flexibility of Markov-switching GAMs with a Beta emission model tailored to bounded responses in (0,1) . Specifically, we introduce a matrix-structured Hidden Markov Model (HMM) to model time series of univariate percentage data. The model features Beta-distributed emissions, where the distribution parameters are governed by a generalized additive model (GAM), thus allowing for smooth dependence on covariates. Each hidden state in the Markov chain determines a distinct Beta distribution, with both the mean (linked via a logistic transformation) and the precision (a scalar) being state-specific. This formulation is particularly suitable for modeling rates, probabilities, or proportions ( 9 ; 6 , see, e.g.,) . Compared to 4 , our model allows the conditional mean of the Beta distribution to vary smoothly with observed covariates within each latent state, while retaining a state-specific precision parameter. Relative to 15 , we specialize the regime-switching additive framework to proportion time series by adopting a Beta mean – precision parametrization, which is natural for rates, shares, and ratios. Methodologically, the contribution is four-fold: (i) We embed a generalized additive model (GAM) with B-spline basis expansions within the Beta emission distribution, allowing the mean to vary nonlinearly across covariates and across latent regimes. (ii) We adopt a frequentist penalized expectation-maximization (EM) estimation strategy rather than a fully Bayesian approach, which is computationally attractive for large samples and avoids the specification of prior distributions for high-dimensional spline coefficient vectors. (iii) We introduce a diagnostic criterion to detect degenerate solutions when the number of latent states is overspecified, and incorporate this criterion into a grid-search-based model selection procedure. (iv) Beyond the algorithmic proposal itself, we provide explicit analytical derivations for the complete-data likelihood and for the derivatives of the penalized \log -likelihood underlying the estimation routine. For this combined Beta-GAM HMM setting, these derivations supply a formal mathematical foundation for the proposed estimation procedure and clarify how the EM updates are constructed. In our proposal, we study proportions that are clearly bounded within the unit interval. Let \{(y_{1},y_{2},\dots,y_{T}),y_{t}\in(0,1)\} be a univariate time series of proportions. For each proportion, we also observe a set of covariates \bm{x}_{t} . To model possibly nonlinear relationships, we let, conditionally on the latent state z_{t}=k , the response follows a Beta distribution with state-specific precision and a mean linked to covariates through a logistic function, y_{t}\mid z_{t}=k\sim\mathrm{Beta}(\mu_{kt}\phi_{k},\,(1-\mu_{kt})\phi_{k}),\qquad\mathrm{logit}(\mu_{kt})=f_{k}(\bm{x}_{t}), where each f_{k}(\cdot) is a smooth function estimated via B-spline basis expansions. Due to its flexibility, we resort to a GAM of 22 and its estimation variants ( 11 , see, e.g.,) . The unknown function f(\cdot) can be estimated by using B -spline functions and their associated coefficients. These coefficients are assumed to be realized with several latent states. In this way, the proposed model can be viewed as a Beta-distributed, covariate-dependent extension of the Beta-HMM in 4 , and at the same time as a proportion-specific specialization of the MS-GAM framework in 15 . We estimate the optimal number of latent states and the smoothing penalty through information criteria combined with diagnostic filtering, and use penalized EM algorithm for estimating the state-dependent mean and precision parameters. The outline of the paper is structured as follows. In Section 2 , we present a HMM framework within the Beta family. The joint distribution of the latent states and the proportions is discussed. In Section 3 , we present the B -spline approximation for the estimation of the unknown function f(\cdot) , where a roughness penalty is placed on the coefficients of the B -splines. In Section 4 , we develop an EM algorithm and a bootstrap to quantifying the uncertainty in parameter estimation. Section 5 introduces diagnostics-based model selection and the associated uncertainty quantification strategy. For completeness, the analytical derivatives that support the penalized estimation routine are collected in Appendix A . Leveraging a series of simulation studies in Section 6 , we investigate the finite-sample properties of the parameter estimation method. Section 7 presents the empirical study investigating a time series of ratios between female and total mortality. Conclusions are given in Section 8 , along with some ideas on how the methodology can be further extended. 2 Hidden Markov Model Framework A HMM provides a flexible probabilistic structure for modeling time series of observed data (y_{1},y_{2},\dots,y_{T}) , assumed to arise from an underlying unobserved sequence of latent states z_{t}\in\{1,2,\ldots,K\} . Each observation may depend on a vector of covariates \bm{x}_{t}\in\mathbb{R}^{p} . The sequence of hidden states (z_{1},z_{2},\ldots,z_{T}) is assumed to evolve according to a first-order Markov chain, with an initial distribution \bm{\pi}=(\pi_{1},\pi_{2},\ldots,\pi_{K})^{\top} , where \pi_{k}=P(z_{1}=k) , satisfying \sum_{k=1}^{K}\pi_{k}=1,\quad\pi_{k}\geq 0. The evolution of the latent state

Read the complete text