Pemodelam Regresi Conway-Maxwell-Poisson untuk Mengatasi Overdispersi pada Data Angka Kematian Ibu di Provinsi Jawa timur
Abstract
Abstract. Poisson regression is usually used to model count data. one assumption in Poisson regression is equidispersion that meanS the mean equals to the variance. However, in real data it is often this assumption does not meet. One way to overcome overdispersion is the Conway-Maxwell-Poisson (COM-Poisson) regression. This study applied the COM-Poisson regression to model the effect of Pregnant women who receive a minimum of 4 antenatal care visits (X1), Active Integrated Health Post (X2), Delivery assisted by Healthcare Professional (X3), Provision of Iron Supplement Tablets to Pregnant Women (X4), Pregnant women who received Td2+ Immunization (X5) and Poverty Rate (X6) to Maternal Mortality Rate (Y) for East Java Province data of 2020. The obtained model is with dispersion parameter 0,35044. Meanwhile, the factors that influence the maternal mortality rate are pregnant women who get at least 4 checkups or check-ups at the end of their pregnancy (X1) and pregnant women who receive health services, especially at posyandu (X2).
Abstrak. Regresi Poisson digunakan dalam memodelkan data cacahan. Salah satu asumsi dalam regresi Poisson adalah equdispersi yang berarti rata-rata sama dengan varians. Namun, pada data rill seringkali asumsi ini tidak terpenuhi. Salah satu cara untuk mengatasi overdispersi adalah dengan menggunakan regresi Conway-Maxwell-Poisson (COM-Poisson). Penelitian regresi COM-Poisson ini diterapkan untuk memodelkan pengaruh dari pemeriksaan akhir masa kehamilan (K4) (X1), keberadaan posyandu aktif (X2), persalinan yang ditolong tenaga kesehatan (X3), pemberian tablet penambah darah pada ibu hamil (X4), pemberian imunisasi td2+ pada ibu hamil (X5) dan persentase penduduk miskin (X6) terhadap Angka Kematian Ibu(Y) di Provinsi Jawa Timur pada data tahun 2020 Diperoleh model Conway-Maxwell-Poisson adalah dengan parameter dispersi 0,35044. Sememtara itu, faktor-faktor yang mempengaruhi angka kematian ibu adalah ibu hamil yang mendapatkan pemeriksaan minimal 4 kali (K4) atau pemeriksaan akhir masa kehamilan (X1) dan ibu hamil yang menerima pelayanan kesehatan terutama di posyandu (X2).
References
Dinas Kesehatan Jawa Timur. (2021). Profil Kesehatan Provinsi Jawa Timur Tahun 2020. Surabaya: Dinkes.
Hilbe, J. M. (2011). Negative binomial regression. Cambridge University Press.
McCullagh, P., & Nelder, J. A. (1989). Generalized linear models. Chapman and Hall. London, UK.
Nelder, J. A., & Wedderburn, R. W. (1972). Generalized linear models. Journal of the Royal Statistical Society Series A: Statistics in Society, 135(3), 370-384.
Sellers, K. F., & Premeaux, B. (2020). Conway–Maxwell–Poisson regression models for dispersed count data. In Wiley Interdisciplinary Reviews: Computational Statistics (Vol. 13, Issue 6). John Wiley and Sons Inc. https://doi.org/10.1002/wics.1533.
Sellers, K. F., & Shmueli, G. (2010). A flexible regression model for count data. The Annals of Applied Statistics, 943-961.
Shmueli, G., Minka, T. P., Kadane, J. B., Borle, S., & Boatwright, P. (2005). A useful distribution for fitting discrete data: revival of the Conway-Maxwell-Poisson distribution. In Appl. Statist (Vol. 54). http://www.blackwellpublishing.com/rss
Suliadi. (2023). Parameter Estimation. Makalah dipresentasikan dalam General Lecture, Program studi Statistika FMIPA Universitas Islam Bandung, Bandung, Bandung.
World Health Organization. (2015). International statistical classification of diseases and related health problems, 10th revision, Fifth edition, Brussel. World Health Organization.
Yang, Z., Hardin, J. W., & Addy, C. L. (2009). A score test for overdispersion in Poisson regression based on the generalized Poisson-2 model. Journal of Statistical Planning and Inference, 139(4), 1514–1521. https://doi.org/10.1016/j.jspi.2008.08.018
Muhammad Rizq Nafisyah Alam, & Aceng Komarudin Mutaqin. (2023). Pemodelan Distribusi Poisson-Sujatha pada Data Frekuensi Klaim Asuransi Kendaraan Bermotor di Indonesia. Jurnal Riset Statistika, 71–78. https://doi.org/10.29313/jrs.v3i1.1944
