基于改进的SAPSO优化支持向量机的太原市空气质量评价

Air Quality Assessment of Taiyuan City Based on Support Vector Machine Optimized by Improved SAPSO

由于气象环境复杂多变且具有动态的不确定特性,选取太原市2014年至2015年的空气污染物监测数据,将模拟退火算法(SA)与粒子群算法(PSO)相结合并对其进行改进,优化支持向量机(SVM)完成参数寻优,并运用偏最小二乘法(PLS)分析各污染物因子间的相互作用,构造出一种新的空气质量评价模型.实验结果表明,改进的SAPSO-SVM与PSO-SVM和SVM相比,模型运行时间短、等级分类精度高,具有良好的评价性能,为空气质量评价提供了新思路.

In view of complex variability and dynamic uncertainty of the atmospheric environment, through the adoption of the monitoring data of air pollutants in Taiyuan city from 2014 to 2015, the paper combined simulated annealing algorithm （SA） with particle swarm optimization algorithm （PSO） and improved it for optimizing support vector machine （SVM） to complete parameter optimization, analyzed the interaction of various pollutants among the factors by using partial least square method （PLS）. A new air quality assessment model was constructed. To compare with PSO-SVM and SVM, the experimental results show that the improved SAPSO-SVM model has the advantages of short running time, higher classification accuracy and a good assessment performance as well as provide a new idea for air quality assessment.