Environmetrics最新文献

In this paper, we propose a Network-Weighted Functional Regression (NWFR) model, an extension of Spatially Weighted Functional Regression (SWFR) to functional data defined on network-structured settings. To assess predictive uncertainty, we develop a functional conformal prediction procedure that yields a distribution-free prediction interval with guaranteed coverage. Through extensive evaluation on both simulated and real-world datasets, we demonstrate that the explicit modeling of network structure yields substantive improvements in point-prediction accuracy and markedly enhances the validity and precision of the resulting prediction intervals.

Estimating parameters of spatial populations requires a sample of response values distributed over the study region. When spatial trends are present, spatially balanced designs give more precise results for commonly used estimators. If auxiliary variables are available, these can also be included in the design to improve precision further. This article proposes a new spatially balanced design to force sample spread in the space of the auxiliary variables. All we require is a distance measure between population units. Numerical results show that the method generates spatially balanced samples and compares favorably with existing designs. We provide two example applications using spatial populations with auxiliary variables and consider equal and unequal probability designs.

Climate change is increasingly recognized as a driver of health-related outcomes, yet its impact on pharmaceutical demand remains largely understudied. As environmental conditions evolve and extreme weather events intensify, anticipating their influence on medical needs is essential for designing resilient healthcare systems. This study examines the relationship between climate variability and the weekly demand for respiratory prescription pharmaceuticals in Greece, based on a dataset spanning seven and a half years (390 weeks). Granger-causality spectra are employed to explore potential causal relationships. Following variable selection, four forecasting models are implemented: Prophet, a Vector Autoregressive model with exogenous variables (VARX), Random Forest with Moving Block Bootstrap (MBB-RF), and Long Short-Term Memory (LSTM) networks. The MBB-RF model achieves the best performance in relative error metrics while providing robust insights through variable importance rankings. The LSTM model outperforms most metrics, highlighting its ability to capture nonlinear dependencies. The VARX model, which includes Prophet-based exogenous inputs, balances interpretability and accuracy, although it is slightly less competitive in overall predictive performance. These findings underscore the added value of climate-sensitive variables in modeling pharmaceutical demand and provide a data-driven foundation for adaptive strategies in healthcare planning under changing environmental conditions.

Samuels, R., N. Carmon, B. Konomi, J. Hobbs, A. Braverman, D. Young, and J. J. Song. 2025. “Estimation of Impact Ranges for Functional Valued Predictors.” Environmetrics 36, no. 5: e70024. https://doi.org/10.1002/env.70024.

In the version of this article initially published, the name of the 3rd author was spelled incorrectly. The correct name is Bledar Konomi, and the spelling error has been updated in the original.

We apologize for this error.

Many environmental processes, such as rainfall, wind, or snowfall, are inherently spatial, and the modeling of extremes has to take into account that feature. In addition, such processes may be associated with a nonextremal feature, for example, wind speed and direction or extreme snowfall and time of occurrence in a year. This article proposes a Bayesian hierarchical model with a conditional independence assumption that aims at modeling simultaneously spatial extremes and an angular component. The proposed model relies on the extreme value theory as well as recent developments for handling directional statistics over a continuous domain. Working within a Bayesian setting, a Gibbs sampler is introduced whose performances are analysed through a simulation study. The paper ends with an application to extreme wind speed in France. Results show that extreme wind events in France are mainly coming from the West, apart from the Mediterranean part of France and the Alps.

Causally-induced asymmetry reflects the principle that an event qualifies as a cause only if its absence would prevent the occurrence of the effect. Thus, uncovering causal effects becomes a matter of comparing a well-defined score in both directions. Motivated by studying causal effects at extreme levels of a multivariate random vector, we propose to construct a model-agnostic causal score relying solely on the assumption of the existence of a max-domain of attraction. Based on a representation of a generalised Pareto random vector, we construct the causal score as the Wasserstein distance between the margins and a well-specified random variable. The proposed methodology is illustrated on a simulated dataset of different characteristics of catchments in Switzerland: discharge, precipitation, snowmelt, temperature, and evapotranspiration.

The block maxima approach, which consists of dividing a series of observations into equal-sized blocks to extract the block maxima, is commonly used for identifying and modeling extreme events using the generalized extreme value (GEV) distribution. In the analysis of coastal wave surge levels, the underlying data that generate the block maxima typically have missing observations. Consequently, the observed block maxima may not correspond to the true block maxima, yielding biased estimates of the GEV distribution parameters. Various parametric modeling procedures are proposed to account for the presence of missing observations under a block maxima framework. The performance of these estimators is compared through an extensive simulation study and illustrated by an analysis of extreme wave surges in Atlantic Canada.