ѧо
RESEARCH OF ENUIRONMENTAL SCIENCES
1999  12 3 vol.12 No.3 1999



ϢҪԵɷַѡ*
Ӿ
ժҪȾֵӳϢʾΪҪԺͽȾҪۺΪȾۺҪԵļ㹫ʽݲϢۺҪԣɷӼѡŻѡ㡣÷˼ҲȷȾۺŻ㡣÷ڹдѡĲܷӳеĻ
ؼʡɷַϢ⡡Żѡ
Optimization of Environmental Monitoring Sites by Principal Component Subset Selection Based on Informative Relative Importance
LI Zuoyong
(Chengdu Institute of Meteorology,Chengdu 610041)
AbstractThe concept of relative importance to the information from the monitoring values of single pollutant was used to select the optimal numbers of environmental monitoring sites,and formula was presented for comprehensive relative importance of multipollutant.On the basis of comprehensive relative importance of information in sites,the optimization of environmental monitoring sites was carried out by principal component subset selection.This method not only reduces the calculation but also determines the comprehensive optimal sites of multipollutant easily.The monitoring sites selected from atmospheric environmental monitoring sites in Guiyang City can represent the atmospheric environmental quality.
KeywordsPrincipal component analysis; Information; Atmospheric environment; Environmental monitoring; Optimal number of sites
ŻеĻϻϣʵķԻɸѡʹֻļݣ׼ȷطӳȾʱշֲͱ仯ɣܵŵӡѧ߷չһЩµĻѡ17ݡɷӼŻѡȡӳĻϢɷ֣ؽѰŷɸѡж֮7ݡַڶȾﻷѡ2㲻㣺ַҪÿȾﶼɷַ󣻢ɲͬȾѡĲȫһ£ԴȷۺϹŻ㡣
ȽĵȾֵʾΪҪϢһƸеԼʽȾҪϢۺϳɶȾۺҪϢȻۺҪϢӦɷӼѡ񷨶Բѡɱѡ
1ϢҪԼۺҪ
1.1ϢҪ
ü񻯹ʽȾֵCijΪCij(±iʱ̣j)

ĳȾֵCijmaxӳϢҪ1CijΪȾֵCijӳϢϢҪԣȻ1Cij0
1.2ۺҪԼ㹫ʽ
ڸȾ֮ϵӣǵϢҪԼȲ໥ų⣬Ҳ໥ˣȾϢۺҪxijӦǸȾϢҪCijļԵӡΪͻCijһȾãۺҪӦxijmaxCij(l),(1lk)kΪȾxijֵmaxxij1ȸеԼʽĶȾϢۺҪԵļ㹫ʽӦΪ

ԸȾ֮໥Ӱ죬ʽ(2)Ϊ

ùʽ(1)(2)ۻʽ(3)ݼʱ̵xij󣬾ͿӦף7ɷӼѡ񷨶Բѡ
2ѡʵ
÷ڶԹ16ѡΪ199212¡19931µ20 dSO2TSPֵͬÿ5վֵԴף8ݡӦùʽ(1)ֱ2Ⱦ12ÿ5վֵݽй񻯴ʽ(2)2ȾۺϺ16㲻ͬʱεۺҪxij1ʾȻ16xijֵطõ16ۺҪϢ֮ϵR1616Ӧף7ɷӼѡ񷨣R1616ֵֵʾϢĹʡÿһɷַ޳1㡣TӦɷӼϷµnѡT޳ĲϢռnϢİٷֱΪ


1㲻ͬʱζȾϢۺҪ

ʱ1ʱ2ʱ3ʱ4
1.Ǿ1.0000.5890.4560.171
2.ˮ೧0.8050.9860.4910.402
3.ҽԺ0.6660.5970.4900.267
4.0.8840.7080.5110.563
5.0.7290.7580.3920.282
6.վ0.8060.8270.4900.236
7.ᷫ0.5940.6100.5550.470
8.ұ0.6260.5070.2930.111
9.·0.8410.8230.5250.661
10.̼ҽ0.8370.9820.7580.597
11.ɹ0.5430.5700.1990.462
12.ؿ0.9860.8120.4920.463
13.ӡ0.6941.0000.5550.414
14.ص0.6330.6610.0950.142
15.С0.6210.8950.5460.443
16.0.3670.5550.2680.063


ʽУnΪnֵСֵµ(n-1)ϢռȫnϢİٷֱΪ

Աʵ̽11ɷӼɸѡ1(Ǿ)6(վ)14(ص)16()4()5㣬5㱣ۺϢռȫ16Ϣ88.73%
3Ƚ
ף8ģԸʵŻѡĵΪ1(Ǿ)6(վ)14(ص)10(̼Ҵ)5()5㣬ˮƽΪ90.6%ʵϣף8ݵľ1610ͬһ2ǳӽĲ㣬ֻ45ͬˣ2ַѡĲ4һµġڲͬķвͬʹͷľޣʶͬһŻѡȫͬ2ַѡĵϢʮֽӽ
4
Ϊ˼ۺҪԵɷӼѡ5Ƿܷӳȫ16ĻҦۺָʽ

16ѡ5ӳĻۺۡʽ(6)CjSjֱΪjȾʵŨȺ͹Ҷ׼mΪȾĿһּ׼2
2һּ׼

ȾȾȾȾ

ָ00.6)0.61.0)1.02.0)2.02.8)2.8)


ֱ16ѡ5ĸȾֵʽ(6)ָI16Ϊ4.02I5Ϊ3.42һּ׼ͬȾ
5
a. жȾŻѡ㷨ֻԶȾۺҪԱɷӼϷѡ㣬Ӷ˼ͲĹ
b. ѡԱ⵱ȾɷӼѡĲ㲻ͬʱȷŻѡ
c. ÷Դ㡢ˮѡʺϣнϺõĿ͹ԡ
*ȻѧĿ(49471048)
߼飺У1944
ߵλɶѧԺɶ610041
6ο
1Venkatram A.Inherent uncertainty in air quality modeling.Atmos Environ,1988,22(6):12211227.
2Grimm J W.Statistical analysis of errors in estimating wet doposition using five surface estimation algorithms.Atmos Environ,1991,25A(2):317320.
3̩̽壬žΰ.ѵ㼯ɸŻ.ѧо19969(1)1215.
4Ӿ.ŻԪģ.̣199614(3)4649.
5Ӿ.Ż˹ģ.йѧ199717(1)2729.
6Ӿ.BP˹Ļѡ.ѧо199811(5)3436.
7ׯᣬ߳ŷٸ.ɷӼѡѡ.йѧ199010(2)148151.
8־.ģŻü̽.ѧչ19964(5)4147.
ոڣ19980915
