This Engineering Doctorate is concerned with change detection. The application of fuzzy clustering and fuzzy classification for online pipeline leak detection is investigated. Three different fuzzy methodologies are derived using one real data set and tested on multiple other real data sets. The first classification methodology, based on a fixed classifier, uses a novel merging algorithm developed to reduce the influence of cluster initialisation on the data space partition. Rules derived from this partition are subsequently used in the second methodology, to define a linguistic rule-based system, classifying without clustering. Instead, it evaluates membership functions to classify. The third methodology is a combination of the previous two, enhanced with cluster rejection and statistical change detection in the form of Shewhart charts.
Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:501952 |
Date | January 2002 |
Creators | Taillefond, D. N. |
Publisher | University of Manchester |
Source Sets | Ethos UK |
Detected Language | English |
Type | Electronic Thesis or Dissertation |
Page generated in 0.0012 seconds