International Journal of applied mathematics and computer science

online read us now

Paper details

Number 4 - December 1996
Volume 6 - 1996

Pattern recognition techniques using fuzzily labeled data for process fault detection

Teodor Marcu

Abstract
The problem of robust diagnosis of process faults is addressed in the context of intelligent control. Fuzzy and non-parametric theoretic decision methods of pattern recognition are developed and applied to model-based fault detection based on parameter estimation. The present approach uses the concept of fuzzy sets to construct and to improve the performances of linear and nonlinear classifiers based on a distance. Simulation and experimental results regarding the diagnosis of various processes are included in two comparative studies.

Keywords
-