最优回归子集法在长爪沙鼠鼠疫疫源地风险分级的应用  

Application of the optimal regression subset on the risk classification of epidemic focus of Meriones unguiculatus plague

作　　者：白雪薇[1] 周松[1] 闫东[1] 牛艳芬[1] 崔耀仁[1] 陈永明[1] 杜国义[1] BAI Xuewei;ZHOU Song;YAN Dong;NIU Yanfen;CUI Yaoren;CHEN Yongming;DU Guoyi(Anti-plague Institute of Hebei Province,Zhangjiakou,Hebei 075000,China)

机构地区：[1]河北省鼠疫防治所,河北张家口075000

出　　处：《医学动物防制》2024年第6期527-530,共4页Journal of Medical Pest Control

基　　金：河北省2018年度医学科学研究重点课题计划(20180956)。

摘　　要：目的分析河北省长爪沙鼠鼠疫疫源地动物鼠疫流行期间的相关危险因素,利用数据指标建立动物鼠疫流行的预测回归模型,对动物鼠疫流行进行风险分级。方法统计河北省长爪沙鼠鼠疫疫源地1990—2013年动物鼠疫监测数据和气象数据,对该鼠疫疫源地动物鼠疫流行总体数据6个指标(长爪沙鼠密度、长爪沙鼠巢蚤指数、长爪沙鼠体蚤指数、极端最低气温、平均相对湿度、日降水量≥0.1mm日数)利用SPSS 26.0软件最优回归子集法进行风险分级模型。将预报值y_(i)>0.50,预报为流行;预报值y_(i)<0.20,预报为不流行;0.20≤y_(i)≤0.50,预报为高风险。结果对风险分级采用实际数据进行拟合,最优回归模型y_(i)>0.50预报流行拟合率为100.00%,y_(i)<0.20预报不流行拟合率为100.00%,0.20≤y_(i)≤0.50预报流行拟合率为71.43%。结论以最优回归子集法模型预报结果为依据建立的河北鼠疫疫源地动物鼠疫流行风险分级模型,拟合效果好、结果准确度高,可为预测动物鼠疫的发生及鼠疫防控策略的制定提供数据支持。Objective To analyze the risk factors associated with animal plague epidemics in the epidemic period focus of Meriones unguiculatus plague in Hebei Province,to establish a predictive regression model of animal plague epidemics using data indicators,and to classify the risk of animal plague epidemics.Methods The monitor data of animal plague epidemics and the meteorological data in the epidemic focus of Meriones unguiculatus plague were counted in Hebei Province from 1990 to 2013.The six indicators(density data of Meriones unguiculatus,nest flea index of Meriones unguiculatus,host flea index of Meriones unguiculatus,extreme minimum temperature,average relative humidity and days with daily precipitation≥0.1mm)of the overall data on the prevalence of animal plague epidemics in the epidemic focus were used to establish a model for the risk classification using the optimal regression subset method in SPSS 26.0 software.When the predicted response y_(i)>0.50,the prediction was epidemic.When the predicted response y_(i)<0.20,the prediction was non-epidemic.When the predicted response 0.20≤y_(i)≤0.50,the prediction was high risk.Results The risk classification was fitted using actual data.The optimal regression model with y_(i)>0.50 predicted the prevalence fitting rate of 100.00%,y_(i)<0.20 predicted non-prevalence fit of 100.00%,and 0.20≤y_(i)≤0.50 predicted prevalence fitting rate of 71.43%.Conclusion Based on the model prediction results of the optimal regression subset method,the risk classification of animal plague epidemics in the epidemic focus in Hebei Province is established with good fitting effect and high accuracy.It can provide data support for the prediction of the occurrence of animal plague and formulation of plague prevention and control strategies.

关 键 词：最优回归子集法 长爪沙鼠鼠疫疫源地 鼠疫流行 动物间 风险分级 河北省 

分 类 号：R181.3[医药卫生—流行病学] R516.8[医药卫生—公共卫生与预防医学]