基于主成分分析和多元线性回归模型的空气质量评价方法研究

发布时间：2018-04-16 15:05

本文选题：环境空气质量 + 规律分析　；参考：《云南大学》2015年硕士论文

【摘要】：环境空气质量是近年来社会关注的热点话题,相关部门公布了实时环境空气质量数据,包括空气质量监测的空气质量指数(AQI)和PM2.5、PM10、SO2、CO、 NO2、O3六项指标浓度等数据。为了从抽象的浓度数值中得出空气质量指数,直观、快速地建立由X(监测指标浓度)到Y(空气质量指数)的映射关系,使其能与现有环境评价方法、法律法规接轨,便于进行环境评价和环境管理。本文采用了趋势图来直观有效的反映年度空气质量指数和六项监测指标之间关系,最后得出监测指标浓度值季节变化规律,从风速、降水、太阳辐射、地理气象因素和污染物排放等方面做了原因分析。之后,本文采用主成分分析法对收集数据进行分析,得出影响空气环境的主要污染物为颗粒物和二氧化氮。再以全年月均值和最高月的日均值代入主成分分析,验证其与空气质量指数实测值变化趋势,预测值和实际值的最高点日期一致、预测值和实际值的最低点日期一致。最后,建立线性回归模型对空气质量指数进行预测,通过多次剔除异常点后,得到多元线性回归模型利用相关系数R2,F检验,P检验,以及下年度的观测值与预测值对比图来验证所得到的线性回归模型与实际数据有较高的相似度。根据规律分析和主要因子计算便于今后更加有针对性地进行环境空气质量管理,数学模型的提出有利于进一步的环境评价和管理。
[Abstract]:The air quality of the environment is a hot topic of social concern in recent years, the relevant departments announced the real-time environmental air quality data, including air quality index of air quality monitoring (AQI) and PM2.5, PM10, SO2, CO, NO2, O3 six index of concentration data. In order to obtain the air quality index, the concentration of numerical Abstract directly, quickly set up by X (monitoring index concentration) to Y (air quality index) and the mapping relation between the existing evaluation methods of environmental laws and regulations, standards for environmental assessment and environmental management. This paper uses the trend diagram showing the relationship between the annual air quality index to effective and six monitoring indicators, finally obtains the monitoring index of concentration variation from seasonal precipitation, wind speed, solar radiation, geological and meteorological factors and pollutant emissions and other aspects of the reason analysis. After that, this paper uses principal component analysis to The data were collected and analyzed, the effects of air pollutants for particulate matter and nitrogen dioxide. The average generation in the component analysis of monthly average and highest monthly, and verify its air quality index measured the change trend, the forecast value and the actual value of the highest point of the date, the forecast value and the actual value is lowest date. Finally, a linear regression model to predict the air quality index, by repeatedly eliminating abnormal points, obtained multiple linear regression model using correlation coefficient R2, F test, P test, and the observation of the annual value and the forecast value comparison chart to verify the linear regression model and the actual data obtained from the high according to the law of similarity. And the main factor analysis calculation for the future more targeted for air quality management, a mathematical model is proposed to further environmental assessment Price and management.

【学位授予单位】：云南大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：X823

【相似文献】