Summary: | 碩士 === 國立交通大學 === 統計所 === 91 === We are interested in estimating the distribution of age of onset. However data of age onset are often censored. In genetic research, the age-onset distribution of a disease may be the penetrance function that represents the cumulative probability of developing a disease by a certain age under a specified genotype. In this thesis, we consider statistical inference based on current status data which is also called as interval censored of case I. Given current status data, we only observe the current age and whether subject has the disease or not but the exact age of onset is never observable. Such data are often seen in epidemiological studies. We consider four methods to estimate the distribution of age-onset given current status data in the presence of long-term survivors. That is, we allow some people to be immune for the disease. Parametric methods include maximum likelihood estimation and two methods based on non-linear regression techniques. At last, we construct nonparametric estimation using self-consistency equation.
|