Extraction of a Parameter as an Index to Assess Water Quality of the Tamagawa, Tokyo, Japan, by Using Neural Networks and Multivariate Analysis

Junko KAMBEa, Tomoo AOYAMAb, Aiko YAMAUCHIc and Umpei NAGASHIMAd*

aFaculty of Foreign Language, Daito Bunka University
1-9-1 Takashimadaita, Itabashi, Tokyo 175-8571, Japan
bFaculty of Technology, Miyazaki University
Gakuen Kihanadai Nishi, Miyazaki, 889-2192, Japan
cGraduate School of Pharmaceutical Sciences, University of Tokushima
1-78 Sho-machi, Tokushima 770-8505, Japan
dResearch Institute of Computational Science, National Institute of Advanced Industrial Science and Technology
1-1-1 Umezono, Tsukuba, Ibaraki 305-8568, Japan

(Received: August 31, 2006; Accepted for publication: October 20, 2006; Published on Web: December 1, 2006)

Parameters as an index to efficiently assess the pollution level of the upper, middle, and lower streams of the Tamagawa (Tokyo, Japan) based on measured water quality were determined by using multivariate analysis, principal component analysis (PCA), and cluster analysis (CA) for data measured from 1994 to 2002. Missing data during 2000-2002 were estimated using a perceptron type neural network and arithmetic progression. The combination of scores for the first and second principal components obtained by PCA enabled classification of the upper, middle, and lower streams of the Tamagawa. The CA results corresponded well with the PCA results.
Based on the score of the first principal compornent determined here, contributions to the water pollution of the middle and lower streams should be decreased to improve the water quality of the Tamagawa.

Keywords: Tamagawa, Principal Component Analysis, Cluster Analysis, Water Contamination, Chemometrics

Text in English

PDF file(125kB)