摘要
为提高水体叶绿素a预测精度和收敛速率,提出一种基于灰色关联度分析和遗传算法优化BP神经网络预测水体叶绿素a的方法。即先采用灰色关联度分析法选取合适的水质指标作为输入因子,然后优化网络隐含层的结构参数,引入遗传算法优化BP神经网络的初始权值和阈值,最后以预测太湖叶绿素a为例进行比较分析。结果表明,优化神经网络隐含层数能进一步提高网络的预测精度、缩短训练时间;灰色关联分析-GA-BP模型相较于BP、GA-BP模型具有更高的预测精度和收敛速度,可为控制水环境监测和决策平台提供科学依据。
In order to improve the accuracy and convergence rate of chlorophyll-aprediction in water body,this paper proposed a forecasting method of chlorophyll-a in water body based on grey relational grade analysis and genetic algorithm to optimize BP neural network.Firstly,suitable water quality indexes were selected as the input factors by grey relational analysis.Then,the structural parameters of the hidden layer of the network were optimized,and genetic algorithm was introduced to optimize the initial weights and thresholds of the BP neural network.Finally,the forecast of chlorophyll-a in Tai Lake was taken as an example for comparative analysis.The results show that optimizing the number of hidden layers of neural network can further improve the prediction accuracy and shorten the training time of neural network.The gray correlation analysis-GA-BP model has higher prediction accuracy and convergence speed than BP and GA-BP model,which can provide scientific basis for controlling water environmental monitoring and decision platform.
作者
朱婕
李翠梅
薛天一
ZHU Jie;LI Cui-mei;XUE Tian-yi(School of Environmental Science and Engineering,Suzhou University of Science and Technology,Suzhou 215009,China;Suning Real Estate Group Corporation,Nanjing 210000,China)
出处
《水电能源科学》
北大核心
2020年第10期25-28,147,共5页
Water Resources and Power
基金
国家自然科学基金项目(51109153)。
关键词
灰色关联法
BP神经网络
遗传算法
叶绿素A
预测
grey relation method
BP neural network
genetic algorithm
chlorophyll-a
prediction