基于因子分析和WOA-Elman的肇庆API预测研究

Prediction of Zhaoqing API Based on Factor Analysis and WOA-Elman Neural Network

为提高肇庆市空气污染指数的预测精度,针对Elman神经网络预测精度受其权值和阈值选择的影响,运用WOA对Elman神经网络的权值和阈值进行优化选择,提出一种融合因子分析和WOA-Elman神经网络的API预测模型.首先结合气候特征对肇庆市春季、夏季、秋季和冬季的API进行月变化特征和季节变化特征分析,然后每种气候下运用因子分析对影响API的多个气象因素进行降维,转换成少数几个因子作为Elman神经网络模型的输入,最后建立基于WOA-Elman神经网络的API预测模型.选择2014年1月1日-2014年12月31日空气质量数据和同期的气象要素资料为研究对象,研究结果表明,算法WOA-Elman能有效提高API的预测精度,具有一定的可行性和指导意义,为肇庆市空气污染防治和政策研究提供科学决策的依据.

In order to improve the prediction accuracy of air pollution index in zhaoqing city,this paper used WOA to optimize the selection of weights and thresholds of Elman neural network,and proposed an API prediction model integrating factor analysis and WOAElman neural network.First combined with the climate features of zhaoqing city spring,summer,autumn and winter API for monthly variation characteristics and seasonal variation characteristics analysis,using factor analysis and then each climate multiple meteorological factors affecting the API for dimension reduction,into a few factors as input of Elman neural network model,the final API based on Elman neural network prediction model is established.Choice on January 1,2014-December 31,2014 air quality data and meteorological data during this period as the research object,the research results show that the PSO-Elman,GA-Elman and Elman,compared this algorithm WOA-Elman can effectively improve the prediction accuracy of API,has certain feasibility and the significance,for zhaoqing city air pollution prevention and policy research to provide the basis for scientific decision.