论文部分内容阅读
采用Monte Carlo模拟仿真方法,研究了仪器分辨率对感度试验方法估值精度的影响,选择常用的3种感度试验方法(升降法、兰利法和D优化法)进行了对比研究。结果表明,仪器分辨率与样本真实标准差的比值,直接影响感度试验的估值精度。当比值小于0.2时,分辨率对感度试验估值精度的影响非常小;当比值大于0.2时,3种感度试验方法有不同的适用性:D优化法适用于比值小于0.4,且对试验样本安全性、可靠性具有较高精度要求的情况;升降法则在比值大于0.4的情况下具有更好的估值精度;兰利法适合于比值小于1.1时,对均值的估值精度要求较高的情况。当比值大于1.1时3种感度试验方法的估值精度都会变低,其中升降法的估值精度最高。
Monte Carlo simulation method is used to study the influence of instrument resolution on the accuracy of the sensitivity test method. Three commonly used sensitivity test methods (lift method, Langley method and D optimization method) are compared. The results show that the ratio of the instrument resolution to the true standard deviation of the samples directly affects the accuracy of the sensitivity test. When the ratio is less than 0.2, the resolution has a very small influence on the accuracy of the sensitivity test. When the ratio is greater than 0.2, the three sensitivity test methods have different applicability: the D-optimization method is suitable for the ratio of less than 0.4 and the safety of the test sample Reliability and reliability with higher accuracy requirements. The lifting rule has a better estimation accuracy when the ratio is greater than 0.4. Langley method is suitable when the ratio is less than 1.1 and the estimation accuracy of the mean value is high . When the ratio is greater than 1.1, the accuracy of the three kinds of sensitivity test methods will be lower, and the accuracy of the lift method is the highest.