RIGHT-CENSORED NEGATIVE BINOMIAL REGRESSION MODEL FOR FERTILITY COUNT DATA AND ITS APPLICATIONS
B. Muniswamy1, Srinu Setti2
1Professor, 2Research Scholar,Department of Statistics, Andhra University, Visakhapatnam-530003, India
Abstract: The research examines the primary determinants of caesarean deliveries in the Indian state of Andhra Pradesh. Estimating characteristics and identifying important factors impacting the number of caesarean section deliveries among women in Andhra Pradesh, India, between the ages of 15 and 49, is the primary goal of the research. The right-censored negative binomial and the right-censored Poisson will be used in the study to achieve this. The fertility count data set, real-world Demographic and Health Surveys phase VII input, and National Family Health Survey, 2019–2021 input are all used in the analysis. Investigating the methods used by expectant mothers to give birth. This study was based on the fitting of the model right-censored number of caesarean section deliveries in right-censored Poisson and right-censored negative binomial utilizing Generalized Additive Models for Location Scale and Shape. The analysis makes use of the R packages survival as well as the add-on package Generalized Additive Models for Location Scale and Shape. The parameters are estimated using both right-censored Poisson and right-censored negative binomial; that are Intercept (-1.7416), “Breech Presentation” (“Yes and Don’t know”) (-0.3211 and 0.5282 ), “Currently has heart disease” (“Yes”) (-0.8697), “High blood pressure” (“Yes and Don’t know”) (0.0323 and 0.0305), “Prolonged labour” (“Yes and Don’t know”) (-0.0707 and -0.6817), “Child is twin” (“1st multiple, 2nd multiple, 3rd multiple”) (0.1565, 0.1288, -8.0569), “Age” (0.0226), “Educational level” (“Primary, Secondary, Higher”) (0.0678, 0.4528, 0.6496) of both right-censored Poisson and right-censored negative binomial. According to the outcomes of applying the Akaike Information Criterion and Schwarz Bayesian Criterion, (4456.177) and (4545.413); (4458.177) and (4553.362) are of right-censored Poisson and right-censored negative binomial, respectively. In the model fitted using right-censored Poisson and right-censored negative binomial, the parameters “Breech presentation” (“Yes and don’t know”), “Prolonged labour” (“Don’t know”), “Age,” and “Educational level” (Secondary and Higher) had an impact on the number of caesarean section deliveries. The government agencies responsible for formulating policies pertaining to women’s welfare must prioritize the health of women, defined as those between the ages of 15 and 49. Compared to the right-censored negative binomial, the right-censored Poisson shows a comparatively better match with number of caesarean section deliveries. To undertake further research, it is recommended to compare right-censored Poisson with several models that predict over-dispersion in count data.
Keywords: count data, caesarean section deliveries, right-censored Poisson, right-censored negative binomial