摘要
针对工业园区废气污染和溯源追踪问题,基于含变异系数滑窗,提出了一种动态关联分析方法.首先,使用变异系数及其变化率来定义滑窗窗口长度的变化,其次,将斯皮尔曼相关系数和最大信息系数加权来计算窗口内两段数据的相关性,反映关联特征,最后在两种方法结合的基础上,使用某工业园区的实际监测数据中的CO浓度数据与其他污染气体浓度进行动态关联分析实验.结果表明,所提出的动态关联分析方法能够快速分析出多种污染废气浓度之间的短时关联特征,为污染预测和溯源提供了重要的参考信息.
Aiming at the problem of waste gas pollution and tracing to the source of industrial park,a dynamic correlation analysis method is proposed based on sliding window with coefficient of variation.Firstly,the variation coefficient and its rate of change are used to define the change of the sliding window length.The Spearman correlation coefficient and the maximum information coefficient are weighted to calculate the correlation between the two segments of data in the window to reflect the correlation characteristics.Finally,based on the combination of the two methods,the CO concentration data in the actual monitoring data of an industrial park is used to carry out the dynamic correlation analysis experiment with the concentration of other pollution gases.The results show that the proposed dynamic correlation analysis method can quickly analyze the short-term correlation characteristics among the concentrations of various pollutants.It provides important reference information for pollution prediction and traceability.
作者
朱涛
王晓凯
卫晓旭
凌德森
ZHU Tao;WANG Xiaokai;WEI Xiaoxu;LING Desen(College of Physical and Electronic Engineering, Shanxi University, Taiyuan 030006, China)
出处
《测试技术学报》
2022年第1期34-41,共8页
Journal of Test and Measurement Technology
关键词
动态关联
斯皮尔曼相关系数
最大信息系数
污染预测和溯源
dynamic association
spearman correlation coefficient
MIC
pollution prediction and traceability