Title | ||
---|---|---|
Statistical Regression Analysis of Threshold Excesses with Systematically Missing Covariates |
Abstract | ||
---|---|---|
AbstractThis work presents a computationally efficient, semiparametric, and nonstationary framework forstatistical regression analysis of threshold excesses with systematically missing covariates basedon the Generalized Pareto Distribution (GPD). The involved regression analysis resolves theinfluence coming from systematically missing covariates by a nonstationary off-set term. Thenonstationarity and the involved ill-posedness of the inverse problem are approached by the FiniteElement time series analysis Methodology with Bounded Variation of model parameters (FEM-BV). Theresulting FEM-BV-GPD approach provides a well-posed problem formulation and goes beyond probabilistica priori assumptions of methods for analysis of extremes based on, e.g., nonstationary Bayesianmixture models and smoothing kernel methods. We compare the performance of the FEM-BV-GPD to thestate-of-the-art approach based on generalized additive models on a test case and on realhistorical precipitation data. |
Year | DOI | Venue |
---|---|---|
2015 | 10.1137/140972184 | Periodicals |
Keywords | Field | DocType |
regression,time series analysis,extreme value theory,generalized Pareto distribution,systematically missing information | Econometrics,Time series,Covariate,Regression analysis,Smoothing,Kernel method,Statistics,Generalized additive model,Mathematics,Mixture model,Bayesian probability | Journal |
Volume | Issue | ISSN |
13 | 2 | 1540-3459 |
Citations | PageRank | References |
0 | 0.34 | 3 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Olga Kaiser | 1 | 0 | 0.34 |
Dimitri Igdalov | 2 | 0 | 0.68 |
Illia Horenko | 3 | 44 | 10.89 |