Sains Malaysiana 52(12)(2023): 3867-3877

http://doi.org/10.17576/jsm-2023-5212-18

 

Estimation of Population Size Based on One-Inflated, Zero-Truncated Count Distribution with Covariate Information

 (Anggaran Saiz Populasi Berdasarkan Taburan Kiraan Satu-Lambung, Sifar-Pemangkasan dengan Maklumat Kovariat)

 

TITA JONGSOMJIT* & RATTANA LERDSUWANSRI

 

Department of Mathematics and Statistics, Faculty of Science and Technology, Thammasat University, Thailand

 

Received: 19 August 2023/Accepted: 12 December 2023

 

Abstract

In order to estimate the unknown size of the population that is difficult or hidden to enumerate, the capture-recapture method is widely used for this purpose. We propose the one-inflated, zero-truncated geometric (OIZTG) model to deal with three important characteristics of some capture–recapture data: zero-truncation, one-inflation, and observed heterogeneity. The OIZTG model is generated by two distinct processes, one from a zero-truncated geometric (ZTG) process, and the other one-count producing process. To explain heterogeneity at an individual level, the OIZTG model provides a simple way to link the covariate information. The new estimator was proposed based on the OIZTG distributions through the modified Horvitz-Thomson approach, and the parameters of the OIZTG distributions are estimated by using a maximum likelihood estimator (MLE). With regard to making inferences about the unknown size of the population, confidence interval estimations are proposed where variance estimate of population size estimator is achieved by using conditional expectation technique. All of these are assessed through simulation studies. The real data sets are provided for understanding the methodologies.

 

Keywords: Capture-recapture; geometric regression; observed heterogeneity

 

AbstraK

Dalam proses untuk menganggarkan saiz populasi yang sukar atau tersembunyi untuk dihitung, kaedah tangkap-tangkap semula digunakan secara meluas untuk tujuan ini. Kami mencadangkan model geometrik satu-lambung, geometrik sifar-pemangkasan (OIZTG) untuk menangani tiga ciri penting bagi beberapa data tangkap-tangkap semula: sifar-pemangkasan, satu-inflasi dan heterogeniti yang diperhatikan. Model OIZTG dijana oleh dua proses yang berbeza, satu daripada proses geometri terpangkas sifar (ZTG) dan satu lagi proses menghasilkan satu kiraan. Untuk menerangkan heterogeniti pada peringkat individu, model OIZTG menyediakan cara mudah untuk memautkan maklumat kovariat. Penganggar baharu telah dicadangkan berdasarkan taburan OIZTG melalui pendekatan Horvitz-Thomson yang diubah suai dan parameter taburan OIZTG dianggarkan dengan menggunakan penganggar kemungkinan maksimum (MLE). Berkenaan dengan membuat inferens tentang saiz populasi yang tidak diketahui, anggaran selang keyakinan dicadangkan dengan anggaran varians penganggar saiz populasi dicapai dengan menggunakan teknik jangkaan bersyarat. Kesemua ini dinilai melalui kajian simulasi. Set data sebenar disediakan untuk memahami metodologi.

 

Kata kunci: Kepelbagaian yang diperhatikan; regresi geometri; tangkap-tangkap semula

 

REFERENCES

Böhning, D. 2008. A simple variance formula for population size estimators by conditioning. Statistical Methodology 5(5): 410–423.

Böhning, D. & Friedl, H. 2021. Population size estimation based upon zero-truncated, one-inflated and sparse count data. Stat. Methods Appl. 30: 1197-1217.

Böhning, D., Kaskasamkul, P. & van der Heijden, P.G.M. 2019. A modification of Chao’s lower bound estimator in the case of one-inflation. Metrika 82: 361-384.

Bunge, J. & Fitzpatrick, M. 1993. Estimating the number of species: A review. Journal of the American Statistical Association 88: 364-373.

Chao, A. & Huggins, R.M. 2006. Four. Modern closed-population capture-recapture models. In Handbook of Capture-Recapture Analysis, edited by Amstrup, S.C., McDonald, T.L. & Manly, B.F.J. Princeton: Princeton University Press. pp. 58-87.

Chao, A. 1987. Estimating the population size for capture-recapture data with unequal catchability. Biometrics 43(4): 783-791.

Godwin, R.T. 2017. One-inflation and unobserved heterogeneity in population size estimation. Biometrical Journal 59(1): 79-93.

Godwin, R.T. & Böhning, D. 2017. Estimation of the population size by using the one-inflated positive Poisson model. Journal of the Royal Statistical Society. Series C (Applied Statistics) 66(2): 425-448.

Good, I.J. 1953. The population frequencies of species and the estimation of population parameters. Biometrika 40(3-4): 237-264.

Horvitz, D.G. & Thompson, D.J. 1952. A generalization of sampling without replacement from a finite universe. Journal of the American Statistical Association 47(260): 663-685.

Lerdsuwansri, R., Sangnawakij, P., Böhning, D., Sansilapin, C., Chaifoo, W., Polonsky, J. & Del Rio Vilas, V. 2022. Sensitivity of contact-tracing for COVID-19 in Thailand: A capture-recapture application. BMC Infectious Diseases 22: 101.

McDonald, S., Hutchinson, S., Schnier, C., McLeod, A. & Goldberg, D. 2014. Estimating the number of injecting drug users in Scotland’s HCV-diagnosed population using capture–recapture methods. Epidemiology & Infection 142(1): 200-207.

Niwitpong, S., Böhning, D., Van Der Heijden, P.G.M. & Holling, H. 2013. Capture–recapture estimation based upon the geometric distribution allowing for heterogeneity. Metrika 76(4): 495-519.

Panyalert, K. & Lanamtaeng, K. 2020. Factors influencing drug rehabilitation attendance at Thunyarak Chiangmai Hospital for substance addiction. Proceedings of the 18th Scientific and Technological Conference, 440-451. Maejo University, Thailand, 28 February 2020.

Tajuddin, R.R.M., Ismail, N. & Ibrahim, K. 2022. Estimating population size of criminals: A new Horvitz–Thompson estimator under one-inflated positive Poisson-Lindley Model. Crime & Delinquency 68(6-7): 1004-1034.

Zelterman, D. 1988. Robust estimation in truncated discrete distributions with application to capture-recapture experiments. Journal of Statistical Planning and Inference 18(2):  225-237.

 

*Corresponding author; email: tita.jong@dome.tu.ac.th

 

 

previous