摘要
本文报告用Monte-Carlo方法对方积乾等提出的职业肿瘤时效分析的数学模型和参数估计方法进行计算机随机模拟的步骤、方法和结果。利用计算机产生2120份单因素单阶段暴露下病例的暴露史和发病时间,以及10010份对照的暴露史,由此形成11组1:4匹配的病例-对照资料和1:2匹配的病例-对照资料,每组含100个病例。对模拟资料分别作加法和乘法模型的参数估计。结果表明,参数的最大似然估计和最大部分似然估计是无偏的,精度满意。
Computer simulation of the mathematical model for time-effect analysisin research of cancer proposed by J.Q.Fang et al and relevant statistical meth-od for parameter estimation is reported in detail.The exposure histories andoccurrence times of disease for 2120 cases and 10010 controls are generatedwith the situation of one risk factor and one-period exposure.1:4 and 1:2matched case-control data are formed from these.The 1:4 matched data aredivided into 11 groups and 1:2 matched ones into 10 groups,each of whichincludes 100 cases.The parameters in additive model and multiplicative modelare estimated on the basis of the above mentioned 1:4 and 1:2 matched datarespectively.The result shows that estimators of patameters based on maximallikelihood and maximal partial likelihood are unbiased and the precision issatisfied.
出处
《中华劳动卫生职业病杂志》
CAS
CSCD
1991年第1期5-7,12,共4页
Chinese Journal of Industrial Hygiene and Occupational Diseases
基金
国家七五攻关项目"适应于肿瘤预防研究的生物统计学理论与技术"
关键词
职业病
肿瘤
预防
时效分析
Time-effect analysis
Computer simulation
Case-control study
