最大似然法在处理检测限以下测量值中的应用  

Application of Maximum Likelihood Estimation in Analysis of the Values Less than Detectable Limit

作　　者：何永华[1] 苗丽壮[1] 傅华[1] 

机构地区：[1]复旦大学医学院预防医学教研室,上海200032

出　　处：《环境与职业医学》2005年第2期151-153,共3页Journal of Environmental and Occupational Medicine

摘　　要：[目的]介绍一种处理检测限以下测量值的方法。[方法]用某炼油厂职业卫生监测数据为例,在EXCEL电子表格中,介绍了应用最大似然法(MLE)处理检测限以下测量值中的步骤。[结果]该现场劳动卫生测定资料服从对数正态分布;MLE均值和标准差无偏估计分别为3 4.96、3 4.5 9mg/m3 ,高于L/2、L/2替代检测限下测量值计算的均值、标准差。[结论]MLE方法合理、简单、实用。Objective] To introduce a new method for treating the values less than measurement detectable limit. [Methods] Steps of maximum likelihood estimation(MLE) in EXCEL spreadsheet were illustrated at length with the example of measurements from a refinery. [Results] The values monitored were of logarithmic normal distribution; unbiased estimations of mean and standard deviation of MLE were 34.96 and 34.59 mg/m3 respectively,they were more than estimations using methods with substitution of L/2 or L/2 for the values less than detectable limit. [Conclusion] MLE is reasonable and easy method for dealing with the values monitored less than detectable limit and may be recommended for practical use.

关 键 词：最大似然法 测量值 检测限 EXCEL电子表格 职业卫生监测 对数正态分布 MG/M^3 劳动卫生 标准差 数据为 炼油厂 均值 

分 类 号：R44[医药卫生—诊断学] R812[医药卫生—临床医学]