Laboratory of Structural Methods of Data Analysis in Predictive
Modeling Moscow Institute of Physics and Technology

Young researchers seminar
Bootstrap tuning in ordered model selection

Описание мероприятия

Владимир Григорьевич Спокойный
21 июня 2016 г. 17:00г.
Москва, Институт проблем передачи информации им. А. А. Харкевича РАН (Б. Каретный пер., 19, метро «Цветной бульвар»), ауд. 615

Даты проведения и дополнительная информация:
Дата Дополнительная информация


In the  problem of model selection for a given family of linear estimators, ordered by their variance, we offer a new "smallest accepted" approach motivated by Lepski's method and multiple testing theory. The procedure selects the smallest model which satisfies an acceptance rule based on comparison with all larger models. The method is completely data-driven and does not use any prior information about the variance structure of the noise: its parameters are adjusted to the underlying possibly heterogeneous noise by the so-called "propagation condition" using a wild bootstrap method.

Дополнительные материалы