Application of Pattern Recognition Techniques to Problems in Advanced Pollution Monitoring.
Abstract
This technical report details the development and implementation of a pattern recognition technique, termed the Fuzzy-c Varieties (FCV) technique. This technique is intended to classify patterns which exhibit membership in only a single class, or data patterns which may partially represent multiple classes. The ability for the patterns to represent multiple classes permits the technique to be of potential value for classifying environmental analysis patterns of mixed samples, such as samples of mixed fuels. The technique was developed and applied to data patterns representing classification measurements on iris flowers, the Fisher iris data set. It was tested further with data patterns representing gas chromatograms of pure and mixed samples of jet fuels. The FCV classification algorithm was implemented as a computer program, written in the FORTRAN computer programming language, and using data display capabilities provided by the X-Windows standard graphical user interface. The technical report describes the classification results obtained by the FCV system from the Fisher iris data set and from the jet fuel data sets. The technical report also provides a user's guide to the FCV computer software.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1995
- Accession Number
- ADA313960
Entities
People
- A. B. Stine
- B. K. Lavine
- X. H. Qin
Organizations
- Clarkson University