论文部分内容阅读
目的在固定样本抽样框架下结合抽样和检测成本探讨基于混合样本的感染率检验中的混合样本大小和检测次数的确定方法。方法利用反正弦变换和正态近似,在控制两类错误概率的前提下建立混合样本量与混合样本大小之间的函数关系,并通过随机模拟考察近似功效函数的近似程度。结果利用数值方法得到不同检验阈值下最小检测次数的整数解及对应的最优混合样本。结论给定两类错误概率,混合样本的大小与检测次数之间此消彼长,实际工作中应根据样品采集成本与检测成本的相对大小加以权衡。
OBJECTIVE: To determine the mixed sample size and the number of detection times in mixed samples based on the sampling rate and detection cost under the framework of fixed sample sampling. Methods Using the arcsine transform and normal approximation, we establish the functional relationship between the mixed sample size and the size of the mixed samples under the premise of controlling the two types of error probabilities, and approximate the approximate power function by stochastic simulation. Results The numerical solution was used to obtain the integer solution of the minimum number of tests under different test thresholds and the corresponding optimal mixed samples. CONCLUSION Given two types of error probabilities, the shift between the size of the mixed sample and the number of tests should be weighed against the relative size of the sample acquisition cost and the test cost in practice.