基于灵敏度和特异度赋权思想的提出对临床诊断的评价提供了更好评价工具的可能，相应的加权Youden指数和加权标准化诊断符合率的方法也已建立。然而，对于比数积、阳性/阴性似然比、阳性/阴性预测值、诊断信息量等更多的评价指标，能否建立一个广义模型来解决所有指标的加权问题却是一个具有广泛应用意义和重要理论价值的课题。本研究将试图建立基于灵敏度与特异度任意赋权的广义加权模型，以彻底解决诊断试验评价指标的加权问题。与此同时，由于诊断试验评价指标众多，哪种方法更优，或哪种方法有什么条件下更优，也是倍受关注但又未很好解决的问题，其主要原因在于现有的模拟研究在构造诊断试验模拟数据时不能与实际数据完全吻合。鉴于此，本研究还将建立检验假设成立但统计量分布未知时基于灵敏度与特异度的参数Bootstrap模拟方法，为诊断试验评价方法的比较奠定坚实基础。

The proposing of the weighted method for sensitivity and specificity provides a better tool to evaluate the diagnostic test, with the weighted Youden index and generalized standardized diagnostic efficiency rate been constructed. However, for other measurements including product odds,diagnostic likelihood ratio, positive/negative predictive value, diagnostic information content, the weighted method remains absence. Can we construct a generalized weighted model which can be applied to all diagnostic test indices? This research has a widely practical and theoretical importance in statistics as well as clinical diagnosis. For this purpose, we will try to construct a generalized weighted model to solve the problem of weighting indices in diagnostic test evaluation completely. In the meantime, because there are so many choices in weighted indices, how to choose a best generalized weighted model or under what conditions is the model more better are two major concerned but not well solved questions. A major cause is that the current simulation methods can not fix the real data in diagnostic test very well. we establish parameter bootstrap method when H0 is true but the distribution of statistic is unknown, which will lay a solid foundation for comparison of different generalized weighted models.