Estimation of Population Size with Heterogeneous Catchability and Behavioural Dependence: Applications to Air and Water Borne Disease Surveillance

TL;DR

This paper introduces a Bayesian trivariate Bernoulli model for population size estimation that accounts for heterogeneity and behavioral dependence, demonstrating robustness and applicability in epidemiological surveillance.

Contribution

A novel Bayesian trivariate Bernoulli model that effectively captures heterogeneity and dependence in capture-recapture data, improving population estimates.

Findings

Model shows robustness under misspecification

Proposed method outperforms existing techniques

Application provides insights into capture-recapture dynamics

Abstract

Population size estimation based on the capture-recapture experiment is an interesting problem in various fields including epidemiology, criminology, demography, etc. In many real-life scenarios, there exists inherent heterogeneity among the individuals and dependency between capture and recapture attempts. A novel trivariate Bernoulli model is considered to incorporate these features, and the Bayesian estimation of the model parameters is suggested using data augmentation. Simulation results show robustness under model misspecification and the superiority of the performance of the proposed method over existing competitors. The method is applied to analyse real case studies on epidemiological surveillance. The results provide interesting insight on the heterogeneity and dependence involved in the capture-recapture mechanism. The methodology proposed can assist in effective decision-making and policy formulation.