arXiv subjects

María Durbán

Publications and source records attributed to María Durbán.

6 recordsLinked to original sources

arXiv2026

Automatic knot selection in smooth additive models

B-spline regression constitutes a widely used framework for nonparametric modeling. The performance of this methodology depends on specifying the number and placement of changepoints, known as knots, prior to the estimation process. Such knot sequence determines the dimension of the B-spline basis used to represent the regression function and the number of coefficients to be estimated. Therefore, the knots' choice affects the model's flexibility, influencing its smoothness and goodness-of-fit. Traditionally, this problem has been addressed either by explicitly selecting knots, via knot-selection algorithms, or by regularization methods, such as P-splines, which automatically tune the regressor's smoothness. The latter have become the standard in generalized additive models (GAMs). In contrast, knot-selection techniques, frequently neglected because of computational or modeling limitations, provide certain advantages which can be valuable in some contexts. In this work, we introduce a novel explicit knot-selection technique for GAMs based on an extension of the adaptive splines (A-splines) knot selection methodology, combined with a customized Fellner-Schall scheme for tuning the associated parameters. Our approach is evaluated on various synthetic and real datasets and compared with P-splines and state-of-the-art knot-selection techniques. The results indicate comparable performance, while producing models built on a substantially smaller number of basis elements.

stat.ML↗

arXiv2026

Variable Domain Multivariate Functional Principal Component Analysis

Multivariate functional principal component analysis (MFPCA) is a powerful dimension reduction technique for analyzing multiple functional variables simultaneously. However, existing MFPCA methods assume that all functional observations are recorded over a common, fixed domain. This assumption is often violated in practical applications where the observation period varies across subjects, leading to what is known as variable domain functional data. We propose a novel approach for MFPCA that explicitly accommodates variable domains by extending existing multivariate functional principal component analysis to the variable domain setting. Our methodology involves performing univariate variable domain FPCA for each functional variable separately, stacking the resulting univariate scores, and then smoothing the empirical covariance matrix of these stacked scores over the domain length. This allows us to estimate multivariate eigenfunctions and scores that properly account for varying observation periods. We demonstrate through extensive simulation studies that our proposed method outperforms approaches that ignore the variable domain structure and rely on binning strategies. The practical utility of our method is illustrated through an application analyzing body temperature and capillary oxygen saturation (SpO$_2$) trajectories in COVID-19 hospital admitted patients, where patients experienced varying lengths of stay and monitoring periods.

stat.ME↗

arXiv2025

Fast Estimation of the Composite Link Model for Multidimensional Grouped Counts

This paper presents a significant advancement in the estimation of the Composite Link Model within a penalized likelihood framework, specifically designed to address indirect observations of grouped count data. While the model is effective in these contexts, its application becomes computationally challenging in large, high-dimensional settings. To overcome this, we propose a reformulated iterative estimation procedure that leverages Generalized Linear Array Models, enabling the disaggregation and smooth estimation of latent distributions in multidimensional data. Through simulation studies and applications to high-dimensional mortality datasets, we demonstrate the model's capability to capture fine-grained patterns while comparing its computational performance to the conventional algorithm. The proposed methodology offers notable improvements in computational speed, storage efficiency, and practical applicability, making it suitable for a wide range of fields in which high-dimensional data are provided in grouped formats.

stat.ME↗

arXiv2020

Multidimensional Adaptive Penalised Splines with Application to Neurons' Activity Studies

P-spline models have achieved great popularity both in statistical and in applied research. A possible drawback of P-spline is that they assume a smooth transition of the covariate effect across its whole domain. In some practical applications, however, it is desirable and needed to adapt smoothness locally to the data, and adaptive P-splines have been suggested. Yet, the extra flexibility afforded by adaptive P-spline models is obtained at the cost of a high computational burden, especially in a multidimensional setting. Furthermore, to the best of our knowledge, the literature lacks proposals for adaptive P-splines in more than two dimensions. Motivated by the need for analysing data derived from experiments conducted to study neurons' activity in the visual cortex, this work presents a novel locally adaptive anisotropic P-spline model in two (e.g., space) and three (space and time) dimensions. Estimation is based on the recently proposed SOP (Separation of Overlapping Precision matrices) method, which provides the speed we look for. The practical performance of the proposal is evaluated through simulations, and comparisons with alternative methods are reported. In addition to the spatio-temporal analysis of the data that motivated this work, we also discuss an application in two dimensions on the absenteeism of workers.

stat.ME↗

arXiv2016

Spatio-temporal adaptive penalized splines with application to Neuroscience

Data analysed here derive from experiments conducted to study neurons' activity in the visual cortex of behaving monkeys. We consider a spatio-temporal adaptive penalized spline (P-spline) approach for modelling the firing rate of visual neurons. To the best of our knowledge, this is the first attempt in the statistical literature for locally adaptive smoothing in three dimensions. Estimation is based on the Separation of Overlapping Penalties (SOP) algorithm, which provides the stability and speed we look for.

stat.ME↗

arXiv2016

Fast estimation of multidimensional adaptive P-spline models

A fast and stable algorithm for estimating multidimensional adaptive P-spline models is presented. We call it as Separation of Overlapping Penalties (SOP) as it is an extension of the \textit{Separation of Anisotropic Penalties} (SAP) algorithm. SAP was originally derived for the estimation of the smoothing parameters of a multidimensional tensor product P-spline model with anisotropic penalties.

stat.ME↗