摘要
针对PM2.5预测中数据来源单一的问题,提出了LSTM融合神经网络预测模型,首先对气候进行特征抽象,作为预测过程的额外输入,与LSTM网络协同分析。以北京市内监测站点2010-2014年间采集的污染数据和天气数据(采样间隔为1小时)进行实验,并将LSTM模型与其他预测模型进行对比,结果显示:该模型相较于LSTM模型RMSE降低了11.6%,MAE降低了14.86%,研究表明本文提出的LSTM融合模型具有更优的预测能力。
the characteristics of climate as an additional part of the prediction process.Input and analyze in collaboration with LSTM network.Experiments with pollution data and weather data(sampling interval of 1 hour)collected from monitoring sites in Beijing from 2010 to 2014,and comparing the LSTM model with other prediction models,the results show that this model is compared to LSTM The RMSE of the model is reduced by11.60%,MAE is reduced by 14.86%.
作者
于书玉
Yu Shuyu(Beijing University of Civil Engineering and Architecture,Beijing,China)
出处
《科学技术创新》
2023年第7期87-90,共4页
Scientific and Technological Innovation
