Chines Journal of Vector Biology and Control ›› 2021, Vol. 32 ›› Issue (3): 339-343.DOI: 10.11853/j.issn.1003.8280.2021.03.016

• Technology and Method • Previous Articles     Next Articles

Application of two different algorithms in prediction of adult mosquito density in Qingdao, China

SONG Fu-cheng1,2, WANG Wei1,2, MA Xiao-fang1,2, LI Bing-hui1,2, XUE Jian-jie1,2, LI Xue-kui1,2, JIANG Hong-rong1,2   

  1. 1. Department of Disinfection and Vector Control, Food Hygiene, Endemic Disease, Qingdao Center for Disease Control and Prevention, Qingdao, Shandong 266033, China;
    2. Qingdao Institute of Preventive Medicine, Qingdao, Shandong 266033, China
  • Received:2020-08-06 Online:2021-06-20 Published:2021-06-20
  • Supported by:
    Supported by the National Science and Technology Major Project of China (No. 2017ZX10303404)

两种不同算法在青岛市成蚊密度预测中的应用

宋富成1,2, 王伟1,2, 马小芳1,2, 李炳辉1,2, 薛建杰1,2, 李学奎1,2, 姜洪荣1,2   

  1. 1. 青岛市疾病预防控制中心消毒与病媒生物防制科/食品卫生科/地方病防制科, 山东 青岛 266033;
    2. 青岛市预防医学研究院, 山东 青岛 266033
  • 通讯作者: 姜洪荣,E-mail:jianghongrong@126.com
  • 作者简介:宋富成,男,硕士,主要从事消毒与病媒生物防制工作,E-mail:qdsongfucheng@126.com
  • 基金资助:
    国家科技重大专项(2017ZX10303404)

Abstract: Objective To predict the density of adult mosquito in Qingdao, China using multiple linear regression algorithm and gene expression programming algorithm models, and to investigate their feasibility in predicting the density of adult mosquito density. Methods Meteorological data (monthly maximum temperature, monthly minimum temperature, monthly mean temperature, monthly mean humidity, monthly accumulated precipitation, and monthly accumulated precipitation days) and monthly adult mosquito density from March to November, 2016-2019, in Qingdao were collected. Prediction models were built using the above two methods, respectively, with the data from 2016 to 2018 used as the training set and the adult mosquito density data in 2019 used as the test set to validate the prediction performance of the two models. Results The correlation coefficients of the training set and the test set were 0.94 and 0.93, respectively, in the multiple linear regression algorithm model, and were 0.97 and 0.96, respectively, in the gene expression programming algorithm model. Conclusion The two models based on the meteorological data can favorably predict the adult mosquito density in Qingdao, which provides data support for mosquito prevention and control in the future.

Key words: Mosquito density, Meteorological data, Multiple linear regression algorithm, Gene expression programming algorithm

摘要: 目的 应用多元线性回归算法和基因表达式编程算法模型预测青岛市成蚊密度,探讨其在成蚊密度预测中的可行性。方法 收集青岛市2016-2019年3-11月的月最高气温、月最低气温、月平均气温、月平均湿度、月累计降水量和月累计降水天数等气象资料和月成蚊密度资料,将2016-2018年数据资料作为训练集分别用2种方法建立预测模型,以2019年成蚊密度数据作为测试集分别验证2种模型的预测性能。结果 多元线性回归算法模型中训练集和测试集的相关系数分别为0.94和0.93。基因表达式编程算法模型中训练集和测试集的相关系数分别为0.97和0.96。结论 基于气象资料建立的2种算法模型均可较好地预测青岛市成蚊密度,为将来开展防蚊灭蚊工作提供了数据支持。

关键词: 成蚊密度, 气象资料, 多元线性回归算法, 基因表达式编程算法

CLC Number: