Water Supply Vol 5 No 3-4 pp 197–208 © IWA Publishing 2005
Identification of key local factors influencing revenue water ratio of Korean cities using principal component analysis and clustering analysis
S. Chung*, H. Lee**, M. Yu***, J. Koo****, I. Hyun***** and H. Lee******
*Department of Environmental Engineering, University of Seoul, 90, Jeonnong-dong, Dongdaemun-gu, Seoul, Korea (E-mail: shinho@ene.uos.ac.kr; hkyoung-lee@hanmail.net; myong@uos.ac.kr; and jykoo@uos.ac.kr
**Department of Environmental Engineering, University of Seoul, 90, Jeonnong-dong, Dongdaemun-gu, Seoul, Korea (E-mail: shinho@ene.uos.ac.kr; hkyoung-lee@hanmail.net; myong@uos.ac.kr; and jykoo@uos.ac.kr
***Department of Environmental Engineering, University of Seoul, 90, Jeonnong-dong, Dongdaemun-gu, Seoul, Korea (E-mail: shinho@ene.uos.ac.kr; hkyoung-lee@hanmail.net; myong@uos.ac.kr; and jykoo@uos.ac.kr
****Department of Environmental Engineering, University of Seoul, 90, Jeonnong-dong, Dongdaemun-gu, Seoul, Korea (E-mail: shinho@ene.uos.ac.kr; hkyoung-lee@hanmail.net; myong@uos.ac.kr; and jykoo@uos.ac.kr
*****Department of Civil and Environmental Engineering, Dankook University, San-8, Hannam-dong, Yongsan-gu, Seoul, Korea (E-mail: ihhyun@dankook.ac.kr
******Water Resources and Environmental Engineering Division, Korea Institute of Construction Technology, 2311, Daehwa-dong, Ilsan-gu, Goyang, Gyonggi-do, Korea (E-mail: hdlee@kict.re.kr
ABSTRACT
In order to identify the relation between revenue water (RW) ratio and key local factors in a quantifiable way, 90 effect factors were considered as regional characteristics for 79 Korean cities. Seven statistically significant effect factors were chosen through correlation analysis. Three principal components independently influencing RW ratio were extracted by principal component analysis (PCA). The 79 cities were grouped into six clusters by k-means clustering (KMC) of the factor scores of the cities. Then key local factors were identified and their impacts were quantified by multiple regression analysis (MRA) and they were justified by T-test and F-test. The approach through correlation-PCA-KMC-MRA was proved to be one of scientific ways for identification of key local factors. According to the result, it was suggested that a shorter length of distribution system, a water supply with smaller number of bigger customer meters a and gravitational supply through reservoir would be advantageous from a RW ratio's point of view. Keywords Key local factors; k-means clustering (KMC); multiple regression analysis (MRA); principal component analysis (PCA); revenue water; water loss management
Full article (PDF Format)
PAY-PER-VIEW: Buy this article for £20.00 (IWA MEMBER PRICE: £15.00)
|
