Title: Data mining of calculations for the control of emissions of organic compounds
Author: B.E.A. Fisher
Author: S. Jensen
Author: P. Waldron
Document Type: Monograph
Annotation: Environmental Policy - Risk and Forecasting. Report No. 27
Abstract:
This report describes the application of data mining techniques for understanding better the relationships between variables within a data set. Data mining techniques may be applied to a wide range of data sets, though they Have mainly found favour within the business community where there may be no underlying scientific explanation for a
relationship between two variables. In this application the results of complex atmospheric chemistry calculations describing the ozone producing potential of volatile organic compounds (VOCs) downwind of a major industrial source has been used as the input data set. This is an interesting application, but was chosen mainly as a convenient data set on which to test the techniques. The full results of the tests have been provided and so are only of value to a specialist reader wishing to gain some insight into the application of data mining techniques, and to see an example o f their use. The full power of the technique may not have been exploited, and there may be further information that could be extracted from the data set. In this example, data mining was not able to reduce the data to a simple, robust form, which would have enabled it to be used for regulatory applications. However it did demonstrate underlying similarities between the behaviour of individual VOCs, which could be understood in terms of their chemical classification e.g. aromatics and alkenes. Data mining is easy to use given the availability of a data mining package. Given the widespread availability of good quality environmental data it would appear that there could be other possibilities for exploiting data mining within the Environment Agency data sets. It may have a role in data reduction, and in exploring underlying relationships, testing data consistency or generating environmental quality criteria. For example successful data mining could result in a simplified classification of VOCs in the Agencys Pollution Inventory.
Publisher: Environment Agency
Publication Date: 2003
Publication Place: Reading
Subject Keywords: Data analysis; Data processing; Data management; Emissions; Modelling; Organic compounds
Extent: 48
Permalink: http://www.environmentdata.org/archive/ealit:492
Total file downloads: 310
Download PDF Display PDF in separate tab
