|
Activity Number:
|
326
- Recent Developments in Probabilistic Record Linkage, Multiple Systems Estimation, and Entity Resolution
|
|
Type:
|
Invited
|
|
Date/Time:
|
Thursday, August 12, 2021 : 10:00 AM to 11:50 AM
|
|
Sponsor:
|
Survey Research Methods Section
|
|
Abstract #316597
|
|
|
Title:
|
Bayesian one-inflated models for population size estimation
|
|
Author(s):
|
Tiziana Tuoto* and Davide Di Cecco and Andrea Tancredi
|
|
Companies:
|
Italian National Institute for Statistics and Sapienza University of Rome and Sapienza University of Rome
|
|
Keywords:
|
record linkage;
capture-recapture;
zero-truncated one-inflated count data;
population size estimation
|
|
Abstract:
|
We present a Bayesian approach to a class of counting models for capture-recapture in presence of "one-inflation". One-inflation has received has received an increasing attention in capture-recapture literature in recent years, particularly in estimating the size of illegal populations. The phenomenon consists in the observation of an excess of individuals captured exactly once. If we do not explicitly model this aspect in the counting distribution, we can overestimate the population size. Bayesian model selection and the role of prior distributions are discussed. Applications to real data for the estimate of the size of some illegal populations are used to illustrate the proposed methodology.
|
Authors who are presenting talks have a * after their name.
