基于遗传小波神经网络的PM_(2.5)浓度预测模型  

PM_(2.5) Concentration Prediction Model Based on Genetic Wavelet Neural Network

作　　者：郑国威 王腾军[1] 杨友森 张翔 ZHENG Guowei;WANG Tengjun;YANG Yousen;ZHANG Xiang(College of Geological Engineering and Geomatics,Chang′an University,Xi′an 710054,China)

机构地区：[1]长安大学地质工程与测绘学院,陕西西安710054

出　　处：《测绘与空间地理信息》2018年第9期248-250,256,共4页Geomatics & Spatial Information Technology

摘　　要：目前,雾霾天气频发,为了提高PM_(2.5)浓度的预测精度,建立了基于遗传算法优化的小波神经网络模型(GA-WNN)。该方法综合了遗传算法的全局搜索能力和小波神经网络(WNN)强大的非线性拟合的优点,弥补了传统神经网络容易陷入局部最小值和收敛速度慢的缺点。以河北省邢台市实时监测的PM_(2.5)浓度数据为样本进行建模预测,预测结果的平均相对误差为11%。将其小波神经网络进行对比分析,实验结果表明:该方法有效地提高了预测精度,为短时的PM_(2.5)含量预测提供了一个新途径。Focusing on frequent haze weather, a wavelet neural network model based on genetic algorithm optimization （GA-WNN） has been established to improve the prediction accuracy in this paper. The method combines the advantages of the global searching ability of genetic algorithm and the strong nonlinear fitting of wavelet neural network （WNN）, which overcomes the shortcomings of the traditional neural network, such as easy to be trapped into local minimum value and slow convergence speed. Based on the real-time monitoring data of PM 2.5 in Xingtai City, Hebei Province.The average relative error of the prediction results is 11%, the comparison and analysis of GA-WNN and WNN show that this method can effectively improve the prediction accuracy and provide a new way for short term PM 2.5 concentration prediction.

关 键 词：遗传算法 小波神经网络 PM 2.5 浓度预测 

分 类 号：P237[天文地球—摄影测量与遥感]