西安市PM_(10)污染预报仿真模型研究  

A PM_(10) Pollution Forecast Simulation Model for Xi'an City

作　　者：涂丽娟[1] 徐仲[1] 张慧清[1] 

机构地区：[1]西北工业大学数学系,陕西西安710072

出　　处：《计算机仿真》2008年第12期110-113,共4页Computer Simulation

基　　金：陕西省自然科学基金(2004CS110002)资助;教育部新世纪优秀人才支持计划资助项目(2006)

摘　　要：为更好的反映西安市环境污染变化,对西安市2004年6月到2006年3月的PM10(可吸入颗粒)污染实测资料与同时期的气象资料进行了分析。首先采用一元线性回归方法,建立PM10污染与各个气象要素在不同季节的相关关系,并得到相关系数,选出对PM10污染影响显著的气象要素;然后采用主成分分析法[2]简化变量;最后用多元线性回归的方法建立回归方程,用最小二乘法[2,5]计算回归系数,并用F检验和t检验对模型的线性性及回归系数的显著性进行检验。最终建立了西安市PM10预报模型体系,计算机仿真结果给出了预报模型命中率及误差。To better reflect the trend of environment pollution of Xi＇an city, this paper analysed the PM10 pollution data and the meteorology data of Xi＇an city from June 2004 to March 2006. The paper first using simple linear regression method, established the relationship between PM10 pollution and various meteorological elements in different seasons, selected elements of larger contribution to the PM10 pollution as prediction factor, simplified the variables by principal component analysis, established equation of regression by multiple linear regression, calculated coefficient of regression by least - squares, verified the linearity and the significance of regression coefficient by F test and t test. Finally, it established the PM10 forecast model system of Xi＇an city, gave the hit rate and error of forecast model.

关 键 词：可吸入颗粒污染 主成分分析 多元线性回归 最小二乘法 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]