A Review of Extreme Value Threshold Estimation and Uncertainty Quantification

By Carl John Scarrott. 2012

539 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/b4b0466861c288ee2c37ecf160fcbf47b7034c80

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Modelling Time Series Extremes (2012, methodology): “In subsequent work Ferro & Segers (2003) proposed the intervals estimator, based on a limiting characterization of the rescaled inter-exceedance intervals: with probability θ an arbitrary exceedance is the last of a cluster, and then the time to the next exceedance has an exponential distribution wi”
Deriving environmental contours from highest density regions (2017, background): “g common parametric sea state joint probability distributions [32], nonparametric models [12] or extreme value models which can have discontinuities at the threshold [29].”
Threshold Selection in Extreme Value Analysis (2016, background): “A recent overview of the topic can be found in Scarrot and MacDonald ([50]).”
Automated threshold selection for extreme value analysis via ordered goodness-of-fit tests with adjustment for false discovery rate (2018, methodology): “Many threshold selection methods are available in the literature; see Scarrott and MacDonald (2012), Caeiro and Gomes (2015) and Langousis et al. (2016) for recent reviews.”
Pareto Smoothed Importance Sampling (2015, methodology): “We tested several other methods reviewed by Scarrot and MacDonald (2012) for selecting u directly, but found them to have higher variability than this simple heuristic, leading to increased variance in estimating k .”
Improved threshold diagnostic plots for extreme value analyses (2014, background): “1.1 Traditional parameter stability plots These plots examine the sensitivity of GP parameter estimates to the threshold and are described in Coles (2001, chapter 4) and Scarrott and MacDonald (2012).”
Space–time modelling of extreme events (2012, background): “The choice of threshold u is typically made informally by using diagnostic plots, sometimes aided by theoretical arguments, though it can be vexing for non-stationary data; see Scarrott and MacDonald (2012) and Northrop and Jonathan (2011) and its discussion.”
Nonstationary frequency analysis of extreme daily precipitation amounts in Southeastern Canada using a peaks-over-threshold approach (2017, background): “The threshold value controls the size of the POT events sample, their independence, and the adequacy of the GPD (Begueria et al. 2011; Coles 2001; Lang et al. 1999; Osman et al. 2013; Scarrott and MacDonald 2012; Zoglat et al. 2014).”
INLA goes extreme: Bayesian tail regression for the estimation of high spatio-temporal quantiles (2018, background): “…to be extreme ; too low a threshold may cause sub-asymptotic bias, while too high a threshold implies large estimation variance due to the small number of threshold exceedances; see Davison and Smith (1990), Northrop and Jonathan (2011) and Scarrott and MacDonald (2012) for standard diagnostics.”