欢迎来到OAJRC平台! 英文版

info@oajrc.org
首页 > 精选文章 > 精选文章

Advances in international applied mathematics. 2020; 2: (1) 1; DOI:10.26855/j.aam.20200001.

Projections of PM2.5 and LSTM Model Construction
PM2.5的预测与LSTM模型构建

作者: 刘东升*

北京市第一六一中学

*通讯作者:刘东升,单位:北京市第一六一中学;

引用本文: 刘东升. PM2.5的预测与LSTM模型构建[J]. 国际应用数学进展, 2020, 2(1) : 1-13.
Published: 2019-12-20

摘要

预测空气质量PM2.5势在必行。影响PM2.5指数的因素很多,很多影响因素具有不确定性和非线性特征,因而对PM2.5预测的效率等都存在一定的问题。 为了监测和估算PM2.5浓度,采用的长短时记忆(LSTM)预测模型能有效地预测空气质量PM2.5指数。为了评估预测PM2.5的LSTM模型的性能,采用了均方误差(MSE)、平均绝对误差(MAE)、均方根误差(RMSE)等测量指标对预测的数据和原始数据进行误差对比分析。通过实验证明,LSTM预测模型能有效地预测空气质量PM2.5指数。本文的主要贡献是根据所获得的污染物浓度因子数据以及地表气象因子数据的不用,在LSTM预测模型基础上,建立了具体的不同数据来源的PM2.5预测模型,从而使PM2.5的预测更具有实用性、可行性和灵活性。

关键词: 北京PM2.5; 预测;LSTM模型

Abstract

Forecasting air quality PM2.5 is imperative. There are many factors that affect the PM2.5 concentration. Many influencing factors have uncertainty and non-linear characteristics, so there are certain problems with the efficiency and accuracy of PM2.5 prediction. In order to monitor and estimate PM2.5 concentration, the LSTM forecasting model can effectively predict the PM2.5 concentration of air quality. In order to evaluate the performance of the LSTM model of PM2.5, the data and the original data are compared and analyzed by measuring indexes such as mean square error(MSE), average absolute error(MAE) and mean square root error(RMSE). Experiments show that the LSTM forecasting model can accurately predict the PM2.5 concentration of air quality. The main contribution of this paper is based on the data of pollutant concentration factor and the use of surface meteorological factor data. Based on the LSTM forecasting model, PM2.5 forecasting model with specific data sources is established. This makes the PM 2.5 forecast more practical, feasible and flexible.

Key words: Beijing PM2.5; Projections; LSTM model

参考文献 References

[1]

[2]

[3]

[4]

[5]

[6]

[7]

  • 274

    总浏览量

浏览了此文章的还浏览了