ESTIMASI PARAMETER REGRESI ZERO-INFLATED NEGATIVE BINOMIAL DENGAN METODE ALGORITMA EXPECTATION MAXIMIZATION (EM) (Studi Kasus: Penyakit Difteri di Jawa Barat Tahun 2016)

Abstract

This study aims to determine the parameter estimation method of the ZINB model with the Expectation Maximization (EM) algorithm, and also to overcome the case of overdispersion caused by excess zeros in diphtheria cases in West Java in 2016 and to know the factors that significantly influence. Based on the results of the analysis, the factors that significantly influence the number of diphtheria cases in West Java Province in 2016 for the data count are the percentage of DPT immunization (X1), the percentage of residents who have access to safe drinking water (X3) and the percentage of TPM that meet sanitary hygiene requirements (X6). Whereas for zero-inflation data is the percentage of residents who have access to decent drinking water (X3), number of puskesmas (X5) and the percentage of TPM that meet sanitary hygiene requirements (X6).