Abstract:A critical problem in extreme value theory (EVT) is the estimation of parameters for the limit probability distributions. Block maxima (BM), an approach in EVT that seeks estimates of parameters of the generalized extreme value distribution (GEV), can be generalized to take into account not just the maximum realization from a given dataset, but the r largest order statistics for a given r. In this work we propose a parameter estimation method that combines the r largest order statistic (r-LOS) extension of BM with permutation bootstrapping: surrogate realizations are obtained by randomly reordering the original data set, and then r-LOS is applied to these shuffled measurements - the mean estimate computed from these surrogate realizations is the desired estimate. We used synthetic observations and real meteorological time series to verify the performance of our method; we found that the combination of r-LOS and bootstrapping resulted in estimates more accurate than when either approach was implemented separately.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to estimate the parameters of the Generalized Extreme Value Distribution (GEV) more accurately in Extreme Value Theory (EVT). Specifically, the paper proposes a new parameter estimation method, which combines the extension of the Block Maxima (BM) - the r - largest order statistics (r - LOS), and the Permutation Bootstrapping. Through this method, the paper aims to improve the accuracy of parameter estimation, especially when there are few extreme values in the data set. ### Background and Motivation In extreme value theory, parameter estimation is a key issue. The traditional Block Maxima (BM) method usually only uses the maximum value of each data block to estimate the parameters of the GEV distribution, while the r - LOS method considers the first r maximum values in each data block, thus using more information. However, both methods have certain limitations in practical applications. In order to further improve the accuracy of parameter estimation, the paper proposes a new method that combines r - LOS and Permutation Bootstrapping. ### Method Overview 1. **r - LOS Method**: - The traditional BM method only uses the maximum value of each data block. - The r - LOS method extends BM and uses the first r maximum values in each data block to estimate the parameters of the GEV distribution. 2. **Permutation Bootstrapping**: - Generate multiple "proxy" data sets by randomly permuting the original data. - Apply the r - LOS method to these proxy data sets and calculate the parameter estimates. - The final parameter estimate is the median of all the proxy data set estimates. ### Experimental Verification The paper verifies the new method through numerical simulation and real - time meteorological time - series data. The results show that the new method has higher accuracy in most cases than using the r - LOS or BM method alone. ### Main Contributions - **Improved the accuracy of parameter estimation**: By combining r - LOS and Permutation Bootstrapping, the new method shows higher accuracy on multiple distributions and data sets. - **Reduced the variability of estimation**: Especially when dealing with meteorological time - series data, the new method significantly reduces the variability of the estimates. - **Provided more robust estimation**: The new method is more robust when dealing with outliers, especially when using the median instead of the mean. In conclusion, the paper proposes a new method that combines r - LOS and Permutation Bootstrapping, aiming to improve the accuracy and robustness of GEV distribution parameter estimation in extreme value theory. Through numerical simulation and real - data verification, the effectiveness of this method has been proven.

Parameter estimation for the generalized extreme value distribution: a method that combines bootstrapping and r largest order statistics

Wild bootstrap inference for penalized quantile regression for longitudinal data

Maximum likelihood estimators based on the block maxima method

Generalized logistic model for r largest order statistics, with hydrological application

Bootstrapping Estimators based on the Block Maxima Method

Gradient boosting for extreme quantile regression

All Block Maxima method for estimating the extreme value index

Fast parameter estimation of Generalized Extreme Value distribution using Neural Networks

EVIboost for the Estimation of Extreme Value Index under Heterogeneous Extremes

Estimation of Extreme Value-at-Risk: an EVT Approach for Quantile GARCH Model

Parameter uncertainty estimation for exponential semi-variogram models: Two generalized bootstrap methods with check- and quantile-based filtering

Functional time series forecasting of extreme values

New Bayesian method for estimation of Value at Risk and Conditional Value at Risk

Estimating quantiles of extreme wind speed using generalized extreme value distribution fitted based on the order statistics

Efficient global estimation of conditional-value-at-risk through stochastic kriging and extreme value theory

Distributionally Robust Optimization as a Scalable Framework to Characterize Extreme Value Distributions

A wee exploration of techniques for risk assessments of extreme events

Bootstrapping for parameter uncertainty in the space–time epidemic-type aftershock sequence model

Fast and scalable inference for spatial extreme value models

Toward Scalable Risk Analysis for Stochastic Systems Using Extreme Value Theory