基于MRMR-SSA-BP的PM_(2.5)浓度预测模型

PM_(2.5) Concentration Prediction Model Based on MRMR-SSA-BP

PM_(2.5)的浓度预测对治理空气和改善环境起着至关重要的作用。以济南市2019年的空气质量数据和气象数据作为研究对象,提出基于最大相关最小冗余算法(MRMR)和麻雀搜索算法(SSA)优化的BP神经网络模型。该模型通过麻雀搜索算法对BP神经网络的初始权值和阈值优化,仿真出BP神经网络初始的最优权值和阈值。把最大相关最小冗余算法选出的最优的特征值作为模型的输入,完成PM_(2.5)浓度的预测。仿真结果表明,与MRMR-BP,SSA-BP,BP等模型相比,MRMR-SSA-BP模型预测效果最佳,为PM_(2.5)浓度的预测提供了一种新的参考方法。

The concentration prediction of PM_(2.5) plays a vital role in controlling the air and improving the environment.Taking the air quality data and meteorological data of Jinan in 2019 as the research object,the BP neural network model was optimized based on the maximum correlation minimum redundancy algorithm(MRMR)and sparrow search algorithm(SSA).The model optimized the initial weight and threshold of BP neural network through sparrow search algorithm,and simulated the initial optimal weight and threshold of BP neural network.The optimal eigenvalue selected by the maximum correlation minimum redundancy algorithm was used as the input of the model to complete the accurate prediction of PM_(2.5) concentration.The simulation results show that compared with MRMR-BP,SSA-BP and BP models,the MRMR-SSA-BP model has the best prediction effect,and provides a new reference method for the prediction of PM_(2.5) concentration.