Water River Pollution Sources Separation Using Independent Component Analysis
| Autor: | Shih-Hsing Yang, 楊士興 |
|---|---|
| Rok vydání: | 2001 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 89 In order to get real independence of the unknown feature patterns, ICA works like the jobs in the processes of data mining (DM), neural networks (NN), machine learning (ML) and blind source separation (BSS). The researches concerning BSS are about speech recognition, data communication, sensor signal processing, and medical science. It’s the same job that ICA want to do, and we have advanced ideas to run ICA in the studies of environmental science. It’s multivariate analysis to solve the problems of the water resources. What I want to do is to find out a new approach to do that in my study here. Independent Component Analysis (ICA) is a new analysis tool. It’s called advanced Principle component Analysis. Computation of covariance is the basis of the PCA to reduce the dimensions of the multivariate analysis. It’s the goal of ICA to get the independence to the unknown patterns of the multivariate data. Sometimes it is difficult to point out what principle components are and what do they mean. I got the USGS water quality data from the internet, including 21 major hydrologic basins of the USA. There’s very good results of running ICA in my study to show out the independence of the IC to be the functions of water quality variables and weights of independent components. We could explain what they are and how they act in the environment by looking into the features of the independent components. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|