arXiv subjects

Ioannis Papastathopoulos

Publications and source records attributed to Ioannis Papastathopoulos.

18 recordsLinked to original sources

arXiv2026

Geometric extremal graphical models and coefficients of extremal dependence on block graphs

We introduce the concept of geometric extremal graphical models, which are defined through the gauge function of the limit set obtained from suitably scaled random vectors in light-tailed margins. For block graphs, we prove results relating to the propagation of various extremal dependence coefficients along the graph. A particular focus is placed on coefficients that link to the framework of conditional extreme value theory, which are especially interesting when variables do not all attain their most extreme values simultaneously. We also consider results related to the case when variables do exhibit joint extreme behaviour. Through the recent translation of the geometric approach for multivariate extremes to a statistical modelling framework, geometric extremal graphical models, and results relating to them, pave the way for an approach to modelling of high dimensional extremes with complex extremal dependence structures.

math.ST↗

arXiv2025

Generative modelling of multivariate geometric extremes using normalising flows

Leveraging the recently emerging geometric approach to multivariate extremes and the flexibility of normalising flows on the hypersphere, we propose a principled deep-learning-based methodology that enables accurate joint tail extrapolation in all directions. We exploit theoretical links between intrinsic model parameters defined as functions on hyperspheres to construct models ranging from high flexibility to parsimony, thereby enabling the efficient modelling of multivariate extremes displaying complex dependence structures in higher dimensions with reasonable sample sizes. We use the generative feature of normalising flows to perform fast probability estimation for arbitrary Borel risk regions via an efficient Monte Carlo integration scheme. The good properties of our estimators are demonstrated via a simulation study in up to ten dimensions. We apply our methodology to the analysis of low and high extremes of wind speeds. In particular, we find that our methodology enables probability estimation for non-trivial extreme events in relation to electricity production via wind turbines and reveals interesting structure in the underlying data.

stat.ME↗

arXiv2025

Statistical inference for radial generalized Pareto distributions and return sets in geometric extremes

We use a functional analogue of the quantile function for probability measures on $\mathbb{R}^d$ to characterize a novel limit Poisson point process for radially recentred and rescaled random vectors under a radial-directional decomposition. This limit process yields new multivariate distributions, including \textit{radial generalised Pareto distributions}, exhibiting stability for extrapolation to extremal sets along any direction. We show that the normalising functions leading to the limit Poisson point process correspond to a novel class of sets visited with fixed probability, with geometric properties determined by the conditional distribution of the radius given the direction and the Radon-Nikodym derivative of the directional probability distribution relative to reference spherical measures. This leads to return sets, defined by the complement of these probability sets and expressed by their return period. We identify an important member, the \textit{isotropic return set}, where all directions of exceedances outside the set are equally likely. Building on the limit Poisson point process likelihood, we develop parsimonious statistical models leveraging links between limit distribution parameters, with novel diagnostics for assessing convergence to the limiting distribution. These models enable Bayesian inference for return sets with arbitrarily large return periods and probabilities of unobserved extreme events, incorporating directional information from observations outside probability sets. The framework supports efficient computations in dimensions d=2 and d=3. We demonstrate the utility of the methods through simulations and case studies involving hydrological and oceanographic data, showcasing potential for robust and interpretable analysis of multivariate extremes.

stat.ME↗

arXiv2023

Bayesian inference of grid cell firing patterns using Poisson point process models with latent oscillatory Gaussian random fields

Questions about information encoded by the brain demand statistical frameworks for inferring relationships between neural firing and features of the world. The landmark discovery of grid cells demonstrates that neurons can represent spatial information through regularly repeating firing fields. However, the influence of covariates may be masked in current statistical models of grid cell activity, which by employing approaches such as discretizing, aggregating and smoothing, are computationally inefficient and do not account for the continuous nature of the physical world. These limitations motivated us to develop likelihood-based procedures for modelling and estimating the firing activity of grid cells conditionally on biologically relevant covariates. Our approach models firing activity using Poisson point processes with latent Gaussian effects, which accommodate persistent inhomogeneous spatial-directional patterns and overdispersion. Inference is performed in a fully Bayesian manner, which allows us to quantify uncertainty. Applying these methods to experimental data, we provide evidence for temporal and local head direction effects on grid firing. Our approaches offer a novel and principled framework for analysis of neural representations of space.

stat.ME↗

arXiv2023

Time series conditional extremes

Accurate modelling of the joint extremal dependence structure within a stationary time series is a challenging problem that is important in many applications.\ Several previous approaches to this problem are only applicable to certain types of extremal dependence in the time series such as asymptotic dependence, or Markov time series of finite order.\ In this paper, we develop statistical methodology for time series extremes based on recent probabilistic results that allow us to flexibly model the decay of a stationary time series after witnessing an extreme event.\ While Markov sequences of finite order are naturally accommodated by our approach, we consider a broader setup, based on the conditional extreme value model, which allows for a wide range of possible dependence structures in the time series.\ We consider inference based on Monte Carlo simulation and derive an upper bound for the variance of a commonly used importance sampler.\ Our methodology is illustrated via estimation of cluster functionals in simulated data and in a time series of daily maximum temperatures from Orleans, France.

stat.ME↗

arXiv2023

Decomposable Tail Graphical Models

We develop an asymptotic theory for extremes in decomposable graphical models by presenting results applicable to a range of extremal dependence types. Specifically, we investigate the weak limit of the distribution of suitably normalised random vectors, conditioning on an extreme component, where the conditional independence relationships of the random vector are described by a chordal graph. Under mild assumptions, the random vector corresponding to the distribution in the weak limit, termed the tail graphical model, inherits the graphical structure of the original chordal graph. Our theory is applicable to a wide range of decomposable graphical models including asymptotically dependent and asymptotically independent graphical models. Additionally, we analyze combinations of copula classes with differing extremal dependence in cases where a normalization in terms of the conditioning variable is not guaranteed by our assumptions. We show that, in a block graph, the distribution of the random vector normalized in terms of the random variables associated with the separators converges weakly to a distribution we term tail noise. In particular, we investigate the limit of the normalized random vectors where the clique distributions belong to two widely used copula classes, the Gaussian copula and the max-stable copula.

math.ST↗

arXiv2023

Hidden tail chains and recurrence equations for dependence parameters associated with extremes of higher-order Markov chains

We derive some key extremal features for $k$th order Markov chains that can be used to understand how the process moves between an extreme state and the body of the process. The chains are studied given that there is an exceedance of a threshold, as the threshold tends to the upper endpoint of the distribution. Unlike previous studies with $k>1$, we consider processes where standard limit theory describes each extreme event as a single observation without any information about the transition to and from the body of the distribution. Our work uses different asymptotic theory which results in non-degenerate limit laws for such processes. We study the extremal properties of the initial distribution and the transition probability kernel of the Markov chain under weak assumptions for broad classes of extremal dependence structures that cover both asymptotically dependent and asymptotically independent Markov chains. For chains with $k>1$, the transition of the chain away from the exceedance involves novel functions of the $k$ previous states, in comparison to just the single value, when $k=1$. This leads to an increase in the complexity of determining the form of this class of functions, their properties and the method of their derivation in applications. We find that it is possible to derive an affine normalization, dependent on the threshold excess, such that non-degenerate limiting behaviour of the process is assured for all lags. These normalization functions have an attractive structure that has parallels to the Yule-Walker equations. Furthermore, the limiting process is always linear in the innovations. We illustrate the results with the study of $k$th order stationary Markov chains with exponential margins based on widely studied families of copula dependence structures.

math.ST↗

arXiv2021

Changes in the distribution of observed annual maximum temperatures in Europe

In this study we consider the problem of detecting and quantifying changes in the distribution of the annual maximum daily maximum temperature (TXx) in a large gridded data set of European daily temperature during the years 1950-2018. Several statistical models are considered, each of which models TXx using a generalized extreme value (GEV) distribution with the GEV parameters varying smoothly over space. In contrast to several previous studies which fit independent GEV models at the grid box level, our models pull information from neighbouring grid boxes for more efficient parameter estimation. The GEV location and scale parameters are allowed to vary in time using the log of atmospheric CO2 as a covariate. Changes are detected most strongly in the GEV location parameter with the TXx distributions generally shifting towards hotter temperatures. Averaged across our spatial domain, the 100-year return level of TXx based on the 2018 climate is approximately 2°C hotter than that based on the 1950 climate. Moreover, also averaging across our spatial domain, the 100-year return level of TXx based on the 1950 climate corresponds approximately to a 6-year return level in the 2018 climate.

stat.AP↗

arXiv2021

Extremal clustering in non-stationary random sequences

It is well known that the distribution of extreme values of strictly stationary sequences differ from those of independent and identically distributed sequences in that extremal clustering may occur. Here we consider non-stationary but identically distributed sequences of random variables subject to suitable long-range dependence restrictions. We find that the limiting distribution of appropriately normalized sample maxima depends on a parameter that measures the average extremal clustering of the sequence. Based on this new representation we derive the asymptotic distribution for the time between consecutive extreme observations and construct moment and likelihood-based estimators for measures of extremal clustering. We specialize our results to random sequences with periodic dependence structure.

math.ST↗

arXiv2020

A spatial Poisson hurdle model with application to wildfires

Modelling wildfire occurrences is important for disaster management including prevention, detection and suppression of large catastrophic events. We present a spatial Poisson hurdle model for exploring geographical variation of monthly counts of wildfire occurrences and apply it to Indonesia and Australia. The model includes two a priori independent spatially structured latent effects that account for residual spatial variation in the probability of wildfire occurrence, and the positive count rate given an occurrence. Inference is provided by empirical Bayes using the Laplace approximation to the marginal posterior which provides fast inference for latent Gaussian models with sparse structures. In both cases, our model matched several empirically known facts about wildfires. We conclude that elevation, percentage tree cover, relative humidity, surface temperature, and the interaction between humidity and temperature to be important predictors of monthly counts of wildfire occurrences. Further, our findings show opposing effects for surface temperature and its interaction with relative humidity.

stat.AP↗

arXiv2018

Testing for equal correlation matrices with application to paired gene expression data

We present a novel method for testing the hypothesis of equality of two correlation matrices using paired high-dimensional datasets. We consider test statistics based on the average of squares, maximum and sum of exceedances of Fisher transform sample correlations and we derive approximate null distributions using asymptotic and non-parametric distributions. Theoretical results on the power of the tests are presented and backed up by a range of simulation experiments. We apply the methodology to a case study of colorectal tumour gene expression data with the aim of discovering biological pathway lists of genes that present significantly different correlation matrices on healthy and tumour samples. We find strong evidence for a large part of the pathway lists correlation matrices to change among the two medical conditions.

stat.ME↗

arXiv2016

Extreme Events of Markov Chains

The extremal behaviour of a Markov chain is typically characterized by its tail chain. For asymptotically dependent Markov chains existing formulations fail to capture the full evolution of the extreme event when the chain moves out of the extreme tail region and for asymptotically independent chains recent results fail to cover well-known asymptotically independent processes such as Markov processes with a Gaussian copula between consecutive values. We use more sophisticated limiting mechanisms that cover a broader class of asymptotically independent processes than current methods, including an extension of the canonical Heffernan-Tawn normalization scheme, and reveal features which existing methods reduce to a degenerate form associated with non-extreme states.

math.PR↗

arXiv2015

Conditional independence and conditioned limit laws

Conditioned limit laws constitute an important and well developed framework of extreme value theory that describe a broad range of extremal dependence forms including asymptotic independence. We explore the assumption of conditional independence of $X_1$ and $X_2$ given $X_0$ and study its implication in the limiting distribution of $(X_1,X_2)$ conditionally on $X_0$ being large. We show that under random norming, conditional independence is always preserved in the conditioned limit law but might fail to do so when the normalisation does not include the precise value of the random variable in the conditioning event.

math.PR↗

arXiv2015

Conditional independence among max-stable laws

Let $X$ be a max-stable random vector with positive continuous density. It is proved that the conditional independence of any collection of disjoint sub-vectors of $X$ given the remaining components implies their joint independence. We conclude that a broad class of tractable max-stable models cannot exhibit an interesting Markov structure.

math.PR↗

arXiv2015

Conditioned limit laws for inverted max-stable processes

Max-stable processes are widely used to model spatial extremes. These processes exhibit asymptotic dependence meaning that the large values of the process can occur simultaneously over space. Recently, inverted max-stable processes have been proposed as an important new class for spatial extremes which are in the domain of attraction of a spatially independent max-stable process but instead they cover the broad class of asymptotic independence. To study the extreme values of such processes we use the conditioned approach to multivariate extremes that characterises the limiting distribution of appropriately normalised random vectors given that at least one of their components is large. The current statistical methods for the conditioned approach are based on a canonical parametric family of location and scale norming functions. We study broad classes of inverted max-stable processes containing processes linked to the widely studied max-stable models of Brown-Resnick, Schlather and Smith, and identify conditions for the normalisations to either belong to the canonical family or not. Despite such differences at an asymptotic level, we show that at practical levels, the canonical model can approximate well the true conditional distributions.

math.PR↗

arXiv2014

Stochastic Ordering under Conditional Modelling of Extreme Values: Drug-Induced Liver Injury

Drug-induced liver injury (DILI) is a major public health issue and of serious concern for the pharmaceutical industry. Early detection of signs of a drug's potential for DILI is vital for pharmaceutical companies' evaluation of new drugs. A combination of extreme values of liver specific variables indicate potential DILI (Hy's Law). We estimate the probability of severe DILI using the Heffernan and Tawn (2004) conditional dependence model which arises naturally in applications where a multidimensional random variable is extreme in at least one component. We extend the current model by including the assumption of stochastically ordered survival curves for different doses in a Phase 3 study.

stat.AP↗

arXiv2012

Dependence Properties of Multivariate Max-Stable Distributions

For an m-dimensional multivariate extreme value distribution there exist 2^{m}-1 exponent measures which are linked and completely characterise the dependence of the distribution and all of its lower dimensional margins. In this paper we generalise the inequalities of Schlather and Tawn (2002) for the sets of extremal coefficients and construct bounds that higher order exponent measures need to satisfy to be consistent with lower order exponent measures. Subsequently we construct nonparametric estimators of the exponent measures which impose, through a likelihood-based procedure, the new dependence constraints and provide an improvement on the unconstrained estimators.

math.ST↗

arXiv2011

Extended Generalised Pareto Models for Tail Estimation

The most popular approach in extreme value statistics is the modelling of threshold exceedances using the asymptotically motivated generalised Pareto distribution. This approach involves the selection of a high threshold above which the model fits the data well. Sometimes, few observations of a measurement process might be recorded in applications and so selecting a high quantile of the sample as the threshold leads to almost no exceedances. In this paper we propose extensions of the generalised Pareto distribution that incorporate an additional shape parameter while keeping the tail behaviour unaffected. The inclusion of this parameter offers additional structure for the main body of the distribution, improves the stability of the modified scale, tail index and return level estimates to threshold choice and allows a lower threshold to be selected. We illustrate the benefits of the proposed models with a simulation study and two case studies.

stat.ME↗