CMStatistics 2021: Start Registration
View Submission - CMStatistics
B1075
Title: Renewal type bootstrap for U- process Markov chains Authors:  Inass Soukarieh - Université de Technologie de Compiègne (France) [presenting]
Abstract: The main purpose is to establish bootstrap uniform functional central limit theorems U-processes for Harris recurrent Markov chains over uniform classes of functions satisfying some entropy condition. To simplify our approach, we will make use of the well-known regenerative properties of Markov chains avoiding some complicated mixing conditions. We show that the U-processes satisfies the bootstrap uniform CLT with minimal condition on envelope function. We next consider an extension to the k Markov chains setting and prove the bootstrap consistency. The theoretical uniform central limit theorems set out are (or will be) key tools for many further developments in Markovian data analysis.