RO EN
Pareto approximation of the tail by local exponential modeling.
Authors: Ion Grama, Vladimir Spokoiny
– 2.25 Mb
Abstract
We give a new adaptive method for selecting the number of upper order statistics used in the estimation of the tail of a distribution function. Our approach is based on approximation by an exponential model. The selection procedure consists in consecutive testing for the hypothesis of homogeneity of the estimated parameter against the change-point alternative. The selected number of upper order statistics corresponds to the first detected change-point. Our main results are non-asymptotic.E-mail: ,
Fulltext
– 2.25 Mb
Contents
- Pareto approximation of the tail by local exponential modeling.
- Classification of GL(2, ℜ)-orbit's dimensions for the differential equations' system with homogeneities of the 4th order.
- Power sets of n-ary quasigroups.
- Existence and uniqueness results for a class of nonlinear differential problems.
- Biharmonic curves in Cartan-Vranceanu (2n+1)-dimensional spaces.
- Infinitely many functional pre-complete classes of formulas in the propositional provability intuitionistic logic.
- Linear convolution of criteria in the vector p-center problem.
- On LCA groups in which some closed subgroups have commutative rings of continuous endomorphisms.
- Multi-dimensional Darboux type differential systems with quadratic nonlinearities.
- GL(2, ℜ) -orbits in a competing species model.
- Minimum Cost Multicommodity Flows in Dynamic Networks and Algorithms for their Finding.
