关于鼠密度监测的贝叶斯计算  

Bayesian computation for rodent density monitoring

作　　者：蒋峥[1] 陈焜 刘朝发 何建邯[1] 廖骏[1] JI ANG Zheng;CHEN Kun;LIU Chao-fa;HE Jian-han;LIAO Jun(Center for Disease Control and Prevention of Chengdu City, Chengdu 610041 , China;Center for Disease Control and Prevention of Longquanyi District, Chengdu 610100, China)

机构地区：[1]成都市疾病预防控制中心,四川成都610041 [2]龙泉驿区疾病预防控制中心,四川成都610100

出　　处：《中华卫生杀虫药械》2019年第2期170-172,共3页Chinese Journal of Hygienic Insecticides and Equipments

摘　　要：目的对2009—2016年成都市某市辖区的鼠密度作贝叶斯计算,以发现监测结果的月间差异。方法按照二项分布贝叶斯模型通过MCMC方法分别计算奇数月与偶数月鼠密度的后验均值。结果偶数月鼠密度后验均值(0.006 428)大于奇数月(0.004 821),并具有统计学差异(后验机会比值1.355)。结论不适宜以单月或双月鼠密度代表全年。Objective To analyze the rodent density in one district of Chengdu form 2009 to 2016 with Bayesian computation for finding inter-monthly difference.Methods Odd and even numbered months rodent density were modeled as binomial distribution respectively by MCMC for posterior means.Results The even numbered months rodent density(the posterior mean was 0.006 428) was greater than odd numbered months(the posterior mean was 0.004 821),there was a statistical significance difference between the two issues(posterior odd ratio was 1.355).Conclusion Odd or even months rodent density is inadequate to represent annual data.

关 键 词：鼠密度 贝叶斯分析 二项分布 WINBUGS 

分 类 号：R184.35[医药卫生—流行病学] S443[医药卫生—公共卫生与预防医学]