Estimating the Density in a Subgroup With Imputed Subgroup Indicators

Research output: Contribution to journalJournal articleResearchpeer-review

26 Downloads (Pure)

Abstract

In this paper, I consider the problem of how to estimate the density in a subgroup when some of the subgroup indicators are missing at random. Four different imputation estimators are compared to each other and to an inverse probability weighted estimator suggested previously. An optimal estimator is derived. I also provide expressions for the asymptotic variance of the imputation estimators including terms of order 1/(nh) and 1/n.
Original languageEnglish
JournalResearch in Statistics
Volume1
Issue number1
Number of pages12
ISSN2768-4520
DOIs
Publication statusPublished - 2023

Keywords

  • Kernel density estimation
  • Imputation
  • Rao-Blackwellization
  • Optimal estimation
  • Variance estimation

Cite this