摘要
精准预测能见度对于保障人们出行质量和交通安全有重要意义,针对目前大气能见度预测算法存在的预测数据来源单一和准确率不高的问题,构建了一个基于遗传算法优化的BP神经网络模型。模型以石家庄市及其卫星城市的6个气象因子和4个主要污染因素(NO2、SO2、PM2.5、PM10)的预处理结果作为输入数据,24 h后的能见度数值作为输出结果。实验结果表明,在弥补BP神经网络模型局部最优问题的基础上,该模型的能见度预测结果在绝对偏差、相关性和预测准确率方面均优于单一采用石家庄气象数据的预测结果,能够提供更为可靠的能见度预报信息。
A BP neural network model based on genetic algorithm is proposed to solve the problem of single source and low accuracy of atmospheric visibility prediction algorithm. The model takes the pre-processing results of six meteorological factors and four main pollution factors(NO2, SO2, PM2.5, PM10) in Shijiazhuang City and its satellite cities as input data, and the visibility value after 24 hours as output data. The experimental results show that the prediction results of the model are better than the single prediction results of Shijiazhuang meteorological data in terms of absolute bias and correlation, which can further improve the prediction accuracy of visibility.
作者
王震洲
聂亚宁
于平平
Wang Zhenzhou;Nie Yaning;Yu Pingping(School of Information Science and Engineering,HeBei University of Science and Technology,Shijiazhuang 050018,China)
出处
《电子测量与仪器学报》
CSCD
北大核心
2019年第11期73-78,共6页
Journal of Electronic Measurement and Instrumentation
基金
河北省科技支撑计划(17210803D)
河北省教育厅青年基金(QN2018095)资助项目
关键词
能见度
遗传算法
BP神经网络
协同预测
visibility
genetic algorithm
BP neural network
collaborative prediction
