基于PSO-LSSVM的区域AQI多维概念预测模型

A Multi-dimensional Conceptual Association Model Based on PSO-LSSVM for Regional AQI

污染物的排放量和气象条件是影响空气质量的主要因素,为了掌握淄博市张店区工业企业对区域环境质量的影响,通过机器学习的方法,基于区内33家重点污染企业的废气日排放量与区域相对湿度、风力等级、平均气温、能见度等气象因素,构建了张店区AQI的PSO-LSSVM多维概念预测模型,反演了张店区2018年1月1日-11月3日的AQI。结果表明:平均气温和AQI有显著的正相关性,相对湿度、能见度与AQI有显著的负相关性;基于PSO-LSSVM建立的AQI预测模型,该预测模型的均方根误差为7.8431,平均绝对误差为6.5037,决定系数为0.8950;该研究可通过气象预报和期望的AQI值得到大气污染物日排放量,为城市环境空气质量保护和企业污染排放调控提供支持。

Emissions of pollutants and meteorological conditions are the main factors affecting the air quality.In order to understand the impact of industrial enterprises on regional environmental quality in Zhangdian District of Zibo City,based on the daily industrial exhaust emissions of 33 key polluting enterprises and meteorological factors such as relative humidity,wind scale,average temperature and visibility in the region,the PSO-LSSVM multi-dimensional conceptual prediction model of AQI in Zhangdian District was established by machine learning method,and the AQI from January 1 to November 3,2018 is inversed.The results showed that there was a significant positive correlation between average air temperature and AQI,while the AQI is negatively correlated with visibility and relative humidity.Based on the AQI prediction model established by PSO-LSSVM,the root mean square error of this prediction model is 7.8431,the average absolute error is 6.5037,and the coefficient of determination is 0.8950.In this study,the daily discharge of atmospheric pollutants can be obtained by meteorological forecast and expected AQI value,which would provide support for urban environmental air quality protection and enterprise pollution emission control.