Estimating a structural distribution function by grouping
Bert van Es, Stamatis Kolios
Abstract
By the method of Poissonization we confirm some existing results concerning consistent estimation of the structural distribution function in the situation of a large number of rare events. Inconsistency of the so called natural estimator is proved. The method of grouping in cells of equal size is investigated and its consistency derived. A bound on the mean squared error is derived.
Create a lesson
Related papers
On Minimum Divergence for Distribution Reconstruction
Justin Bec, Thomas Le Corre, Aurélien Garivier
Estimators from Convex Geometries of the Simplex
Paul Navas
Two Sample Sizes in Test-Time LLM Ensembling: Buffered ERM and Minimax Theory
Zhi Zhang, Arash A. Amini
A One-Sided McDiarmid-Type Inequality for Heavy Tails via Reverse KL Tensorization
Hanuka Eshima, Amedeo Roberto Esposito, Andrea Paudice
Factor-Assisted Density Estimation: When Ambient Dimension Enables Oracle Rates
Sunrit Chakraborty, David Dunson
Precise Asymptotics for Dense High-Dimensional Huber Regression under Heteroscedasticity
Jing Zhou, Ramji Venkataramanan, Hui Zou