Search

Quantifying the risks of extreme scenarios requires understanding the tail behaviours of variables of interest. While the tails of individual variables can be characterized parametrically, the extremal dependence across variables can be complex and its modeling remains one of the core problems in extreme value analysis. Notably, existing measures for extremal dependence, such as angular components and spectral random vectors, reside on nonlinear supports, such that statistical models and methods designed for linear vector spaces cannot be readily applied. In this paper, we show that the extremal dependence of $d$ asymptotically dependent variables can be characterized by a class of random vectors residing on a $(d-1)$-dimensional hyperplane. This translates the analyses of multivariate extremes to that on a linear vector space, opening up the potentials for the application of existing statistical techniques, particularly in statistical learning and dimension reduction. As an example, we show that a lower-dimensional approximation of multivariate extremes can be achieved through principal component analysis on the hyperplane. Additionally, through this framework, the widely used H\"usler-Reiss family for modelling extremes is characterized by the Gaussian family residing on the hyperplane, thereby justifying its status as the Gaussian counterpart for extremes.

In this paper, we estimate the sparse dependence structure in the tail region of a multivariate random vector, potentially of high dimension. The tail dependence is modeled via a graphical model for extremes embedded in the H\"usler-Reiss distribution. We propose the extreme graphical lasso procedure to estimate the sparsity in the tail dependence, similar to the Gaussian graphical lasso in high dimensional statistics. We prove its consistency in identifying the graph structure and estimating model parameters. The efficiency and accuracy of the proposed method are illustrated by simulations and real data examples.

In an earthquake event, the combination of a strong mainshock and damaging aftershocks is often the cause of severe structural damages and/or high death tolls. The objective of this paper is to provide estimation for the probability of such extreme events where the mainshock and the largest aftershocks exceed certain thresholds. Two approaches are illustrated and compared -- a parametric approach based on previously observed stochastic laws in earthquake data, and a non-parametric approach based on bivariate extreme value theory. We analyze the earthquake data from the North Anatolian Fault Zone (NAFZ) in Turkey during 1965-2018 and show that the two approaches provide unifying results.

The $k$-means clustering algorithm and its variant, the spherical $k$-means clustering, are among the most important and popular methods in unsupervised learning and pattern detection. In this paper, we explore how the spherical $k$-means algorithm can be applied in the analysis of only the extremal observations from a data set. By making use of multivariate extreme value analysis we show how it can be adopted to find "prototypes" of extremal dependence and we derive a consistency result for our suggested estimator. In the special case of max-linear models we show furthermore that our procedure provides an alternative way of statistical inference for this class of models. Finally, we provide data examples which show that our method is able to find relevant patterns in extremal observations and allows us to classify extremal events.

In many statistical modeling frameworks, goodness-of-fit tests are typically administered to the estimated residuals. In the time series setting, whiteness of the residuals is assessed using the sample autocorrelation function. For many time series models, especially those used for financial time series, the key assumption on the residuals is that they are in fact independent and not just uncorrelated. In this paper, we apply the auto-distance covariance function (ADCV) to evaluate the serial dependence of the estimated residuals. Distance covariance can discriminate between dependence and independence of two random vectors. The limit behavior of the test statistic based on the ADCV is derived for a general class of time series models. One of the key aspects in this theory is adjusting for the dependence that arises due to parameter estimation. This adjustment has essentially the same form regardless of the model specification. We illustrate the results in simulated examples.

The world's urban population is expected to grow fifty percent by the year 2050 and exceed six billion. The major challenges confronting cities, such as sustainability, safety, and equality, will depend on the infrastructure developed to accommodate the increase. Urban planners have long debated the consequences of vertical expansion---the concentration of residents by constructing tall buildings---over horizontal expansion---the dispersal of residents by extending urban boundaries. Yet relatively little work has predicted the vertical expansion of cities and quantified the likelihood and therefore urgency of these consequences. We regard tall buildings as random exceedances over a threshold and use extreme value theory to forecast the skyscrapers that will dominate the urban skyline in 2050 if present trends continue. We predict forty-one thousand skyscrapers will surpass 150 meters and 40 floors, an increase of eight percent a year, far outpacing the expected urban population growth of two percent a year. The typical tall skyscraper will not be noticeably taller, and the tallest will likely exceed one thousand meters but not one mile. If a mile-high skyscraper is constructed, it will hold fewer occupants than many of the mile-highs currently designed. We predict roughly three-quarters the number of floors of the Mile-High Tower, two-thirds of Next Tokyo's Sky Mile Tower, and half the floors of Frank Lloyd Wright's The Illinois---three prominent plans for a mile-high skyscraper. However, the relationship between floor and height will vary across cities.

Preferential attachment is an appealing edge generating mechanism for modeling social networks. It provides both an intuitive description of network growth and an explanation for the observed power laws in degree distributions. However, there are often limitations in fitting parametric network models to data due to the complex nature of real-world networks. In this paper, we consider a semi-parametric estimation approach by looking at only the nodes with large in- or out-degrees of the network. This method examines the tail behavior of both the marginal and joint degree distributions and is based on extreme value theory. We compare it with the existing parametric approaches and demonstrate how it can provide more robust estimates of parameters associated with the network when the data are corrupted or when the model is misspecified., Comment: 23 pages, 6 figures, 2 tables

Regular variation is often used as the starting point for modeling multivariate heavy-tailed data. A random vector is regularly varying if and only if its radial part $R$ is regularly varying and is asymptotically independent of the angular part $\Theta$ as $R$ goes to infinity. The conditional limiting distribution of $\Theta$ given $R$ is large characterizes the tail dependence of the random vector and hence its estimation is the primary goal of applications. A typical strategy is to look at the angular components of the data for which the radial parts exceed some threshold. While a large class of methods has been proposed to model the angular distribution from these exceedances, the choice of threshold has been scarcely discussed in the literature. In this paper, we describe a procedure for choosing the threshold by formally testing the independence of $R$ and $\Theta$ using a measure of dependence called distance covariance. We generalize the limit theorem for distance covariance to our unique setting and propose an algorithm which selects the threshold for $R$. This algorithm incorporates a subsampling scheme that is also applicable to weakly dependent data. Moreover, it avoids the heavy computation in the calculation of the distance covariance, a typical limitation for this measure. The performance of our method is illustrated on both simulated and real data., Comment: 21 pages, 6 figures

Preferential attachment is an appealing mechanism for modeling power-law behavior of the degree distributions in directed social networks. In this paper, we consider methods for fitting a 5-parameter linear preferential model to network data under two data scenarios. In the case where full history of the network formation is given, we derive the maximum likelihood estimator of the parameters and show that it is strongly consistent and asymptotically normal. In the case where only a single-time snapshot of the network is available, we propose an estimation method which combines method of moments with an approximation to the likelihood. The resulting estimator is also strongly consistent and performs quite well compared to the MLE estimator. We illustrate both estimation procedures through simulated data, and explore the usage of this model in a real data example., Comment: 34 pages, 6 figures

The use of empirical characteristic functions for inference problems, including estimation in some special parametric settings and testing for goodness of fit, has a long history dating back to the 70s (see for example, Feuerverger and Mureika (1977), Csorgo (1981a,1981b,1981c), Feuerverger (1993)). More recently, there has been renewed interest in using empirical characteristic functions in other inference settings. The distance covariance and correlation, developed by Szekely and Rizzo (2009) for measuring dependence and testing independence between two random vectors, are perhaps the best known illustrations of this. We apply these ideas to stationary univariate and multivariate time series to measure lagged auto- and cross-dependence in a time series. Assuming strong mixing, we establish the relevant asymptotic theory for the sample auto- and cross-distance correlation functions. We also apply the auto-distance correlation function (ADCF) to the residuals of an autoregressive processes as a test of goodness of fit. Under the null that an autoregressive model is true, the limit distribution of the empirical ADCF can differ markedly from the corresponding one based on an iid sequence. We illustrate the use of the empirical auto- and cross-distance correlation functions for testing dependence and cross-dependence of time series in a variety of different contexts., Comment: 28 pages, 6 figures

For the directed edge preferential attachment network growth model studied by Bollobas et al. (2003) and Krapivsky and Redner (2001), we prove that the joint distribution of in-degree and out-degree has jointly regularly varying tails. Typically the marginal tails of the in-degree distribution and the out-degree distribution have different regular variation indices and so the joint regular variation is non-standard. Only marginal regular variation has been previously established for this distribution in the cases where the marginal tail indices are different.