DOI: 10.5176/2251-1911_CMCGS16.30
Authors: Oyelola A. Adegboye
Abstract: This study presents the use of finite mixture for analyzing epileptic seizure counts which allows for modeling inhomogeneous populations with a different probability density function in each component. Epilepsy is a disease that is often misunderstood thus leading to fear, secrecy, stigmatization and the risk of social and legal penalties. It is widely characterized by the spontaneous and unforeseeable occurrence of seizures during which the perception or behavior of patients is disturbed. A finite mixture has a finite number of components; it offers a simple and natural model for unobserved population heterogeneity under the condition that number of unobserved subpopulation is fixed. In this study thirteen component Poisson mixtures were obtained and patients were classified into different sub-populations. There was no patient classified into the seventh sub-population. Poisson regression indicates significant interaction between baseline rate and sup-population, and treatment and suppopulation.
Keywords: component; finite mixture; Poisson; epileptic seizure
Updating...