摘要
为了提高空气质量预测精度,提出一种由GentleAdaBoost(GAB)迭代算法和模糊BP神经网络组成的空气质量预测方法.首先,对空气质量样本数据进行预处理,并初始化测试数据分布权值;然后,通过选取不同类型的模糊BP神经网络构造出不同的模糊BP弱预测器,并对样本数据进行反复训练;最后,将多个模糊BP神经网络弱预测器使用GAB算法进行迭代,得到新的强预测器.对北京市近1a的空气污染指数(API)和空气质量指数(AQI)进行仿真实验,结果表明:本方法比传统BP空气质量预测模型预测平均误差绝对值减少近30%,提高了神经网络预测精度,为神经网络在空气质量预测中的应用提供借鉴.
In order to improve the accuracy of air quality forecasting,a new method combining the GentleAdaBoost(GAB)iterative algorithms with fuzzy BP(back propagation)neural network was put forward.Firstly,the method performed the pretreatment for the air quality historical data and initialized the distribution weights of test data.Secondly,it selected different types of fuzzy BP neural network and trained the sample data repeatedly.At last,multiple weak predictors of fuzzy BP neural network were iterated by GAB algorithm and formed a new strong predictor.Some simulation experiments for the database of air pollution index(API)and air quality index(AQI)in Beijing were carried out.The results show that this method has reduced the average error value by more than 30%compared to the traditional BP air quality forecasting model,and has improved the forecasting accuracy of neural network.This method provides references for the air quality forecasting.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2013年第S1期63-65,69,共4页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家星火计划资助项目(2011GA690190)
江苏省属高校自然科学重大基础研究资助项目(11KJA460001)
关键词
神经网络
迭代算法
模糊理论
空气质量
预测
neural network
iterative algorithms
fuzzy theory
air quality
forecasting
