Plenary lecture 1: integrating fuzzy regression and fuzzy time series into knowledge-based modelling
Anna Walaszek-Babiszewska · 2009
The regression methods and time series models have a long history of applications to data analysis and statistical inference dealing with randomness of the modelling systems. Modelling techniques based on fuzzy sets, especially fuzzy rule based models are suitable for modelling nonlinear and complex systems with imprecise information concerning the system structure and variables. The rules can be seen as local submodels of the system, operating in particular fuzzy regions. In this work we present the methods of the creating the probabilistic-fuzzy knowledge base and the inference procedure. The rules have fuzzy antecedents representing linguistic values of input variables and consequents in the form of fuzzy regression equations or fuzzy time series models. This approach can be seen as an extension of the concept of Takagi-Sugeno fuzzy models. Such knowledge base and inference procedure, as parts of an expert system, are capable of supporting human decisions in control, prediction, diagnosis and in many fields of activity.