• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 1
  • Tagged with
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

Pattern recognition based on qualitative representation of signals. Application to situation assessment of dynamic systems

Gamero Argüello, Fco. Ignacio (Francisco Ignacio) 26 June 2012 (has links)
The main focus of situation assessment is to decide on the adequacy of process behaviour with respect to specifications. When is not possible to have a mathematical model to represent the system operation, other non-model-based techniques must be considered. Classification methods are typically proposed as strategies for diagnosis. Here, identification of the functional states is reduced to recognising the current shapes of variables as well-known states, commonly taking advantage of a process expert or past experiences. However, human knowledge is related to concepts and symbols whereas process acquisition systems provide monitoring systems with numerical data. Consequently, these type of knowledge-based decision systems are usually forced to work in a higher level of abstraction using symbolic representations. This thesis deals with the study of classification methods when performing qualitative trends analysis. The aim is to obtain qualitative trends and their classification by means of the extracted knowledge from past experiences. This doctoral dissertation deals with the study of classification methods when performing qualitative trends analysis. The aim is to obtain qualitative trends and their classification by means of the extracted knowledge from past experiences. / El objetivo principal de la evaluación de situaciones es decidir sobre la adecuación del comportamiento del proceso con respecto a las especificaciones. Cuando no es posible tener un modelo matemático para representar el funcionamiento del sistema, otras técnicas deben considerarse. Los métodos de clasificación suelen ser propuestos como estrategias para el diagnóstico. La identificación de los estados funcionales se reduce a reconocer las formas de las variables como estados conocidos, comúnmente adquiriendo conocimiento de un experto o experiencias anteriores. Sin embargo, el conocimiento humano se relaciona con conceptos y símbolos, mientras que los sistemas de adquisición proporcionan datos numéricos. En consecuencia, este tipo de sistemas basados en el conocimiento de decisiones trabajan en un nivel superior de abstracción a través de representaciones simbólicas. Esta tesis aborda el estudio de métodos de clasificación de las tendencias cualitativas. El objetivo es clasificarlas por medio del conocimiento extraído de las experiencias pasadas.

Page generated in 0.0774 seconds