多变量灰色模型MGM(1，n)在白纹伊蚊密度预测中的应用  

Application of multivariable grey model( 1,n) in prediction of aedes albopictus density

作　　者：黄建华[1] 石挺丽 陈远源 陈少威[1] 张宗昀 尹嘉熙 陈清[1] 俞守义[1] 

机构地区：[1]南方医科大学公共卫生与热带医学学院流行病学教研室,广东广州510515 [2]南方医科大学公共卫生与热带医学学院预防医学专业,广东广州510515

出　　处：《中华疾病控制杂志》2016年第1期87-90,共4页Chinese Journal of Disease Control & Prevention

基　　金：广州市医药卫生重大科技项目(2006-Zda-001)

摘　　要：目的利用伊蚊诱捕器监测法得出指标与气象资料建立多变量灰色预测模型（1，n）（multivariablegreymodel，MGM（1，n）），对伊蚊密度进行短期预测。方法以广州市某农村为研究现场，使用幼虫监测法和伊蚊诱捕器监测法监测伊蚊密度，并收集同期气象资料。利用2014年7～11月的诱蚊诱卵指数（mosquitoandoviposi-tionpositiveindex，MOI）和5项气象资料指标与布雷图指数（breteauindex，BI）进行灰色关联度分析，选择关联度较大的变量建立MGM（1，n）灰色预测模型，使用12月资料验证模型预测效果。结果各指标与BI的灰色关联序为MOI，相对湿度，最高平均气温，降雨量，平均气温和最低平均气温。采用BI与MOI建立MGM（1，2）模型，BI拟合值和实测值的平均绝对误差为9．14，平均相对误差为34．73％。而对MOI的拟合值和实测值的平均绝对误差为2．04，平均相对误差为21．44％。对12月伊蚊密度进行预测，BI预测值与实测值平均绝对误差为1．23，而MOI平均绝对误差1．43。结论多变量灰色预测模型MGM（1，2）能对白纹伊蚊密度进行短期的预测。Objective To make a short-term prediction about aedes albopictus density by muhivariable grey model （MGM） （ 1 ,n）, based on the aedes trap monitoring data and climate data. Methods Field investigation on aedes albop- ictus density at a village in Guangzhou was performed using aedes trap monitoring and traditional larval monitoring, and me- teorological data were collected simultaneously. MGM （1, n） was established based on the variables with maximum slope degree of grey incidence, which was calculated by grey relational analysis utilizing mosquito and oviposition positive index （MOI） , and five items of meteorological data and breteau index（BI） , from July to November, 2014. The data in Decem- ber was used to verify the prediction effect of the model. Results MOI had the maximum slope degree of grey incidence, then relative humidity, maximum mean temperature, precipitation, mean temperature and minimum mean temperature fol- lowed. MGM （1,2） was developed by BI and MOI. The average absolute errors were 9. 14 and 2. 04 for BI and MOI, re- spectively. The average relative errors were 34. 73% and 21.44%. When forecasting the density of aedes in December, the average absolute errors were 1. 23 and 1.43 for BI and MOI, respectively. Conclusions MGM （1,2） can be used to make a short-term prediction for aedes albopictus density.

关 键 词：伊蚊属 模型 统计学 流行病学方法 

分 类 号：R181[医药卫生—流行病学]