基于气象参数的某城市空气质量二次预测模型

A Secondary Prediction Model of Air Quality in a City Based on Meteorological Parameters

摘要: 本文主要研究了空气质量预报问题，广泛收集了各类相关数据，剔除异常数据并利用平均差值补全缺失数据得到计算数据。对气象参数进行合理分类，并计算相关系数。通过超限学习机神经网络建立预测模型预测各污染物的浓度，在一次预测的基础上采用遗传算法对模型的参数进行优化和再预测，使预测出的空气质量指数(AQI)的相对误差最大值最小，首要污染物误差最小。通过对比分析预测值验证了二次模型的合理性。

Abstract: This paper mainly studies the problem of air quality forecasting, collects various relevant data ex-tensively, removes abnormal data, and uses the average difference to complete the missing data to obtain the calculated data. The meteorological parameters are reasonably classified and the corre-lation coefficients are calculated. A prediction model is established to predict the concentration of each pollutant through an ELM neural network, and the genetic algorithm is used to optimize and re-predict the parameters of the model on the basis of one prediction, so as to maximize the relative error of the predicted air quality index (AQI). The smallest value is the smallest primary pollutant error. The reasonableness of the quadratic model is verified by comparing and analyzing the pre-dicted values.