Your browser does not support JavaScript!
2016-12-19(一),主講人:黃世豪 博士 (Department of Statistics, University of Michigan)


統 計 學 研 究 所

專 題 演 講


講 題: Budget constrained group testing design for prevalence estimation with an imperfect assay and a gold standard

黃世豪博士 (Department of Statistics, University of Michigan)

時 間: 105年12月19日(星期一)13:20 - 14:20  (12:00 - 13:20 茶會於統計所821室舉行)
地 點: 綜合三館837室
摘 要:

We consider optimal group testing design when a cheap imperfect assay and an expensive perfect (gold standard) assay for the target trait are both available. For large-scale usage, the primary assay for detecting a trait is usually imperfect, and its sensitivity and specificity may vary among different population. Even if there exists a gold standard assay with no testing error, it may be too expansive or time-consuming for routine use. Moreover, the cost of testing a group tends to increases with group size. The primary goal is to accurately estimate the prevalence of the trait in a given population, where the sensitivity and specificity of the imperfect assay are treated as nuisance parameters. We use budget constraints to reflect the costs of performing either of the two assays relative to the cost of collecting an individual sample. Intuitively, a mixed design strategy should be adopted, where each sample is tested by either only the imperfect assay, only the perfect assay, or both assays. We prove that the optimal budgeted designs have at most three group sizes using only the imperfect assay, one using only the perfect assay, and one using both assays. Moreover, we provide an algorithm to obtain an optimal budgeted design.

This is a joint work with Prof. Mong-Na Lo Huang and Prof. Kerby Shedden.


敬請公佈                                 歡迎參加

  • 友善列印
  • 分享
Voice Play