摘要
为了降低PM2.5设备预测成本,同时分析大气因素与污染物的相关性,选取O3、CO、PM10、SO2、NO2污染物指标预测PM2.5,之后加入温度、湿度、风力等大气指标,建立综合气象的指标体系,采用pearson算法对指标进行合并,用pearson相关指标的BP神经网络的方法再次对PM2.5做预测。对比实验证明,基于pearson相关指标的BP神经网络PM2.5预测模型在提高了预测准确率的同时降低了预测的时间复杂度,起到了降低PM2.5预测成本的目的。
In order to reduce the prediction cost of PM2. 5 equipment , and to analyse the cor relat ion be-tween atmospheric factors and pollutants, the pollutant index of O 3 、CO 、PM10、 SO2 、 NO 2 was selected to predict PM2. 5,then added temperature,humidity,wind power and other atmospheric indexes,for establis-hing the comprehensive meteorological index system. Using Pearson algorithm to merge the index, with the method of BP neural network related indicators again on PM2. 5 forecast. Compared with the experi-mental results, the BP neural network model based on Pearson related index in PM2. 5 can improve the prediction accuracy and reduce the time complexity of prediction, which can reduce the prediction cost of PM2. 5.
出处
《青岛大学学报(自然科学版)》
CAS
2017年第2期83-87,91,共6页
Journal of Qingdao University(Natural Science Edition)
基金
安徽省高校自然科学重点项目(批准号:KJ2015A309)资助