基于湿度分区的线性拟合臭氧浓度预测方法  

An Ozone Concentration Prediction Method Based on Humidity Partition Linear Fitting

作　　者：刘建晓[1,2] 王悦[3] 孟令辉[1] 魏淑珍[2] LIU Jianxiao;WANG Yue;MENG Linghui;WEI Shuzhen(College of Electronics&Information Engineering,Hengshui University,Hengshui,Hebei 053000,China;Hebei Collaborative Innovation Center for Wetland Conservation and Green Development,Hengshui University,Hengshui,Hebei 053000,China;College of Life Science,Hengshui University,Hengshui,Hebei 053000,China)

机构地区：[1]衡水学院电子信息工程学院,河北衡水053000 [2]衡水学院河北省湿地保护与绿色发展协同创新中心,河北衡水053000 [3]衡水学院生命科学学院,河北衡水053000

出　　处：《衡水学院学报》2023年第4期13-19,77,共8页Journal of Hengshui University

基　　金：衡水市科技计划项目(2021014006Z);衡水学院院级项目(2021ZR27)。

摘　　要：为获取当地臭氧与气象条件的关系,根据衡水市臭氧浓度变化特征和衡水超级站实测数据,提出一种基于湿度分区线性拟合的臭氧浓度预测方法。根据历史实测数据,建立了一年前的臭氧-湿度散点图。由于臭氧浓度分布不符合高斯分布特征,可将湿度区间划分为若干个线性区域,在每个线性区域对臭氧数据进行多元函数拟合,建立臭氧数据与气象条件的线性关系,比较测试样本数据和分区数据的欧式距离,按照欧式距离的最小值对样本数据进行分区。将建立的分区模型应用于样本数据集,得到样本数据个数对应的臭氧浓度。通过气象条件分析,最终选择温度和湿度两个变量作为预测条件,取得了良好的预测效果。In order to obtain the relationship between local ozone and meteorological conditions,an ozone concentration prediction method based on humidity partition linear fitting is proposed according to the change characteristics of ozone concentration in Hengshui City and the measured data of Hengshui super station.A year-old ozone-humidity scatter plot is established based on the historical measured data.Since the ozone concentration distribution does not conform to the Gaussian distribution,the humidity interval can be divided into several linear regions,and the ozone data is fitted with a multivariate function in each linear region to establish a linear regression model between the ozone data and the meteorological conditions.In addition,the Euclidean distance between the test sample data and the partition data is compared,and the sample data is partitioned according to the minimum value of the Euclidean distance.The established partition model is applied to the sample data set to obtain the ozone concentration corresponding to the number of sample data.Through the analysis of meteorological conditions,two variables,temperature and humidity,are finally selected as the prediction conditions,and a good prediction effect is achieved.

关 键 词：臭氧 湿度 分区 线性回归 预测 

分 类 号：X820.4[环境科学与工程—环境工程] X513