University of British Columbia
Weak convergence of empirical processes with network dependent variables
Abstract
dc:descriptionThis paper derives a weak convergence to a Gaussian process result for empirical processes indexed by a general function class satisfying a certain bracketing entropy condition in the network 𝜓-dependence setting as introduced in Kojevnikov et al. (2021) with modifications to incorporate multisets. The covariance structure of the limiting Gaussian process is given, and asymptotic equicontinuity of empirical processes is shown as a consequence. An analogous weak convergence result in the time series 𝜓-dependence setting is implied by this paper's result. A Bernstein-type inequality is proven as an important intermediate result, which is also applied to obtain a standard uniform convergence rate of a kernel density estimator with compactedly supported, Lipschitz functions to the sample average density function of heterogeneous multivariate network dependent data. A key condition supporting these results shows a tradeoff between the incremental growth of network neighbourhoods and the decay rate of dependence strength in the network. This paper's results provide tools for a range of econometric applications with network dependent observations including semiparametric estimation with nuisance, M-estimation, and nonstandard testing.
Degree
thesis:*- Name thesis:degree_name
- Master of Arts - MA
- Level thesis:degree_level
- master's
- Discipline thesis:degree_discipline
- Economics
- Grantor dc:publisher
- University of British Columbia
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Li, Chenyu
- Contributors dc:contributor
-
- Song, Kyungchul
Rights
dc:rights- Statement dc:rights
-
- Attribution-NonCommercial-NoDerivatives 4.0 International
- Language dc:language
- eng
Identifiers
dc:identifier.*- Handle dc:identifier
- http://hdl.handle.net/2429/95296