Spatial Modeling Weibull-3 Survival Parameters with Frailty Distributed Conditionally Autoregressive (CAR)

Nur Mahmudah (Program Studi Statistika Jurusan Matematika Fakultas MIPA Universitas Brawijaya Malang)
Henny Pramoedyo (Jurusan Matematika Fakultas MIPA Universitas Brawijaya Malang)

Natural B
Vol 3, No 1 (2015)

Survival analysis is a collection of statistical procedures for data analyzing, where respon variables caused by time until an event occurs. One of application of survival regression’s purpose is to know dengue hemorragic fever. Since the spread of dengue hemorragic fever caused by the spread of mosquito, there is probability that event in one location affects other event in another locations thus, it is better to model with Bayessian method of spatial survival. Model includes random spatial effect CAR to overcome the spatial effect in survival model using queen contiguity type weight. This study aimed to obtain spatial survival model one survival data year of 2013 which was the event of dengue hemorragic fever in city of Malang. Based on the data, moran value I was -0.5930 with Z-test value equal to -2,002, which means there is a spatial autocorelation on the event of dengue hemorragic fever in city of Malang. Spatial survival model with Weibull-3 Parameter (Weibull-3P) distribution obtained the factors significantly affecting dengue hemorragic fever, which were sex, hematrocit rate, thrombocyte volume had equal rate of healing in each subdistrict.