PENANGANAN OVERDISPERSI MENGGUNAKAN REGRESI GENERALIZED POISSON DAN BINOMIAL NEGATIF PADA KASUS KEMATIAN IBU HAMIL AKIBAT HIPERTENSI

RAHADATUL AISY NABILAH, . (2023) PENANGANAN OVERDISPERSI MENGGUNAKAN REGRESI GENERALIZED POISSON DAN BINOMIAL NEGATIF PADA KASUS KEMATIAN IBU HAMIL AKIBAT HIPERTENSI. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.

[img] Text
COVER.pdf

Download (939kB)
[img] Text
BAB 1.pdf

Download (355kB)
[img] Text
BAB 2.pdf
Restricted to Registered users only

Download (845kB) | Request a copy
[img] Text
BAB 3.pdf
Restricted to Registered users only

Download (507kB) | Request a copy
[img] Text
BAB 4.pdf
Restricted to Registered users only

Download (651kB) | Request a copy
[img] Text
BAB 5.pdf
Restricted to Registered users only

Download (414kB) | Request a copy
[img] Text
DAFTAR PUSTAKA.pdf

Download (359kB)
[img] Text
LAMPIRAN.pdf
Restricted to Registered users only

Download (530kB) | Request a copy

Abstract

Jumlah kematian ibu merupakan salah satu data bertipe cacahan. Salah satu sebaran peluang yang umum digunakan untuk mendeskripsikan data bertipe cacahan adalah sebaran Poisson. Karakteristik dari sebaran Poisson adalah kondisi equidispersi, yaitu kondisi dimana nilai ragam sama dengan rataannya. Namun berdasarkan fakta di lapangan, seringkali terjadi pelanggaran asumsi equidispersi, yaitu terjadi kondisi dimana nilai ragam yang diamati lebih besar daripada nilai rataannya (overdispersi). Overdispersi dapat menyebabkan standard error pada pendugaan parameter menjadi underestimate. Untuk mengatasi kondisi overdispersi, pada penelitian ini akan digunakan model regresi Generalized Poissondan Binomial Negatif yang bertujuan untuk mengidentifikasi faktor-faktor yang signifikan mempengaruhi jumlah kasus kematian ibu hamil akibat hipertensi di Indonesia berdasarkan model terpilih. Hasil penelitian menunjukkan bahwa model regresi Generalized Poisson dan Binomial Negatif relatif mampu mengatasi overdispersi karena menghasilkan nilai dispersi yang mendekati satu. Sementara itu, berdasarkan kriteria evaluasi model dengan AIC, AICc, dan BIC, model regresi Binomial Negatif memiliki performa yang paling baik dibandingkan kedua model lainnya. Model regresi Binomial Negatif menghasilkan model yang lebih sederhana dengan tiga variabel penjelas yang signifikan mempengaruhi jumlah kasus kematian ibu hamil akibat hipertensi di Indonesia, diantaranya persentase cakupan pelayanan kesehatan ibu hamil K4, persentase cakupan puskesmas yang menyediakan kelas ibu hamil, dan persentase penduduk miskin. **** The number of maternal deaths is a count type data. One of the probability distributions commonly used to describe count-type data is the Poisson distribution. The characteristic of the Poisson distribution is the equidispersion condition, which is the condition where the variance is equal to the mean. However, based on the facts in the field, there are often violations of the equidispersion assumption, namely conditions where the observed variance is greater than the average value (overdispersion). Overdispersion can cause the standard error in parameter estimation to underestimate. To overcome the overdispersion condition, this study will use Generalized Poisson and Negative Binomial regression models which aim to identify significant factors that influence the number of cases of maternal deaths due to hypertension in Indonesia based on the selected model. The results showed that the Generalized Poisson and Negative Binomial regression models were relatively able to overcome overdispersion because they produced dispersion values close to one. Meanwhile, based on the model evaluation criteria with AIC, AICc, and BIC, the Negative Binomial regression model has the best performance compared to the other two models. The negative binomial regression model produces a simpler model with three explanatory variables that significantly influence the number of cases of maternal deaths due to hypertension in Indonesia, including the percentage of coverage for K4 pregnant women, the percentage of coverage of public health center providing classes for pregnant women, and the percentage of poor people.

Item Type: Thesis (Sarjana)
Additional Information: 1). Dr. Dian Handayani, M.Si. ; 2). Vera Maya Santi, M.Si.
Subjects: Sains > Sains, Ilmu Pengetahuan Alam
Sains > Matematika
Sains > Matematika > Software, Sistem Informasi Komputer
Divisions: FMIPA > S1 Statistika
Depositing User: Users 20541 not found.
Date Deposited: 11 Sep 2023 06:21
Last Modified: 11 Sep 2023 06:21
URI: http://repository.unj.ac.id/id/eprint/42027

Actions (login required)

View Item View Item