Activity Number:
|
479
- Applied Bayesian Nonparametric Modeling
|
Type:
|
Contributed
|
Date/Time:
|
Thursday, August 6, 2020 : 10:00 AM to 2:00 PM
|
Sponsor:
|
Section on Bayesian Statistical Science
|
Abstract #313736
|
|
Title:
|
Nonparametric Bayesian Population Size Estimation with Missing Entries
|
Author(s):
|
Dongah Kim* and Krista Gile
|
Companies:
|
University of Massachusetts, Amherst and University of Massachusetts, Amherst
|
Keywords:
|
Nonparametric;
Bayesian;
Capture-Recapture;
missing values;
mixture model
|
Abstract:
|
We propose a Bayesian nonparametric method for estimating closed population size based on multiple lists with missing entries. Based on the Bayesian population size estimation developed by \cite{Manrique2016}, we extended this approach when the missing entries exist. This approach, based on Dirichlet process mixture models, can capture the complex dependencies without assuming a specific number of latent classes. We deal with missing entries by using data-augmentation, jointly fitting Bayesian Capture-Recapture models and modeling the missing data process. We apply this new method in a simulation study with several scenarios and to data on casualties in the Syrian civil war.
|
Authors who are presenting talks have a * after their name.