Title: Further Development and Testing Of Artificial Intelligence Systems for the Classification and Diagnosis of River Quality Based on Biological and Environmental Data
Author: M.A. O'Connor
Author: Environment Agency
Author: M. F Paisley
Author: D.J. Trigg
Author: W.J Walley
Document Type: Monograph
Annotation: Environment Agency Project ID:EAPRJOUT_1008, Representation ID: 299, Object ID: 2289;
Environment Agency Project ID:EAPRJOUT_1323, Representation ID: 436, Object ID: 2504
Abstract:
This report outlines the results of an extension to R&D Project E1-056 (Walley et al., 2002), which was financed by the Environment Agency. The aims were to: - test the computer-based systems River Pollution Diagnostic System (RPDS) and River Pollution Bayesian Belief Network (RPBBN) developed in E1-056 using data derived from the newly acquired 2000 general quality assessment (GQA) survey; - update and improve RPDS by including a means of identifying and incorporating ‘reference states’ and a methodology to extend RPDS to produce a classification system; - retrain RPDS using number of families in place of alkalinity in the input vector; - produce an updated database to be used as input to the existing RPDS; - identify means for updating and improving the current prototype RPBBN. As proposed in the original specification of the extension, no changes were made to the existing RPDS and RPBBN software, as it would require considerable time, cost and effort to update the systems on the Environment Agency’s network. Instead, the proposed changes remain largely theoretical. A new database has been produced for RPDS, including a retrained model with both 1995 and 2000 data. Although the database is not complete (e.g. it does not contain RIVPACS classifications of sites from the 2000 National River Quality Survey of England and Wales, or figures for feeding group composition at each site), it contains sufficient information to be used with the standard version of RPDS.
Publisher: Environment Agency
Subject Keywords: Rivers; Invertebrates; Water pollution; Water quality; Environmental factors; Computer science; Quality
Extent: 39
Permalink: http://www.environmentdata.org/archive/ealit:4538
Total file downloads: 321
Download PDF Display PDF in separate tab
