Улучшенная процедура выбора модели для оценивания диффузионных сигналов по дискретным данным С. С. Перелевский
Material type: ArticleContent type: Текст Media type: электронный Other title: Improved model selection procedure for diffusion signal processing from discrete data [Parallel title]Subject(s): численное моделирование | задачи оценивания | метод наименьших квадратовGenre/Form: статьи в сборниках Online resources: Click here to access online In: Перспективы развития фундаментальных наук. Т. 3 : сборник научных трудов XVIII Международной конференции студентов, аспирантов и молодых ученых, 27–30 апреля 2021 г Т. 3 : Математика. С. 70-72Abstract: In this paper we consider the nonparametric estimation problem for the diffusion signal on the basis of discrete data with unknown diffusion coefficients. For this problem through the truncated sequential point wise kernel estimators we construct a new improved model selection procedure for which we show oracle inequalities for arbitrary orthonormal basis on the grid, i.e. we show, that the constructed model selection procedure is optimal in the sharp oracle inequalities sense.Библиогр.: 3 назв.
In this paper we consider the nonparametric estimation problem for the diffusion signal on the basis of discrete data with unknown diffusion coefficients. For this problem through the truncated sequential point wise kernel estimators we construct a new improved model selection procedure for which we show oracle inequalities for arbitrary orthonormal basis on the grid, i.e. we show, that the constructed model selection procedure is optimal in the sharp oracle inequalities sense.
There are no comments on this title.