| Summary: | 碩士 === 國立成功大學 === 統計學系碩博士班 === 95 === The analysis of the population size estimation has a long history and has been widely used in many applications. In this research field, it usually can be divided into the open population and the closed population. Under the closed population, we often assumed capture history of animals is the completely data. However, in the practical applications, animals maybe lose marks during sampling periods, records mistake of the marks or records lost. It causes the capture history of animals incomplete. Therefore, this paper studies the estimators of population size for the closed population. The bias of the estimators under complete capture-recapture data and capture history with missing data are compared. Two biases will be compared when capture history data are generated with missing completely at random and the probability that capture history data is missing depends on the capture-recapture data value. Finally, a bias corrected estimator is proposed.
|
|---|
