컨텐츠 시작
학술대회/행사
초록검색
| 제출번호(No.) | 0585 |
|---|---|
| 분류(Section) | Invited Talk |
| 분과(Session) | Probability / Stochastic Process / Statistics (SS-12) |
| 영문제목 (Title(Eng.)) |
Finite population distribution function estimation using P-splines |
| 저자(Author(s)) |
Sumanta Adhya1, Tathagata Bandyopadhyay2, Gourangadeb Chatterjee3 West Bengal State University, India1, Indian Institute of Management Ahmedabad, India2, University of Calcutta, India3 |
| 초록본문(Abstract) | Estimation of finite population distribution function (hereafter FPDF) is an important problem in survey sampling since it summarizes almost all the relevant information about the finite population characteristics of interest. Chambers and Dunstan (1986) (henceforth CD) develop a predictive estimator of FPDF using linear regression model in the super-population to incorporate unit level information in the population on a set of auxiliary variables. Here we extend CD’s estimator using a semiparametric regression model in the super-population based on recently developed penalized splines (P-splines) regression and find it’s asymptotic bias and variance. The performance of the proposed estimator is compared with the competing model-based estimators through a model-based simulation study. We also propose a bootstrap hybrid estimator of its variance as discussed in the previous chapters and prove its consistency when simple random sampling design is used. The performance of the proposed variance estimator is compared with that of the analytical estimators through a simulation study. Finally, a real data set is used to compare the performance of the proposed estimator with the competing estimators, both model-based and design-based, through a design-based simulation study. |
| 분류기호 (MSC number(s)) |
62D05 |
| 키워드(Keyword(s)) | finite population distribution function, penalized splines, bootstrap estimation |
| 강연 형태 (Language of Session (Talk)) |
English |