A generalized method for developing partially formalized objects identification fuzzy models by direct rule generation based on experimental data is formulated. Models built according to this principle have the intrinsic ability to operate in accordance with the observation data. Under the condition of the initial experimental data set being representational enough, they may not even require additional tuning of the membership functions parameters. Still, systems developed based on experimental data are often redundant, and may require corrections of the input feature set magnitude. An approach for modifying the number of model inputs is proposed. It allows to do so without the model losing its capability to adequately reflect the subject area. In order to develop a fuzzy logic system, which would reflect the subject area in an adequate manner, an optimization criterion is proposed, measuring the increase in mutual information reflecting from a fuzzy logic system’s inputs to its outputs. Under the condition of maintaining the system’s capability for adequate decision making, a sequence of steps required for developing a type-2 fuzzy logic system, optimal according to the considered criterion, is shown. This paper provides justification for type-2 fuzzy sets being appropriate for use in mathematical models dealing with uncertain input data. The justification is performed theoretically, based on information theory considerations, and confirmed experimentally.The proposed method enables solving applied problems of identifying multidimensional objects, such as an environmental system Author Biographies Natalia Kondratenko, Vinnytsia National Technical University Khmelnytske highway, 95, Vinnytsia, Ukraine, 21021 PhD, Associate Professor Department of information security
Alan : Fen Bilimleri ve Matematik
Dergi Türü : Uluslararası
Benzer Makaleler | Yazar | # |
---|
Makale | Yazar | # |
---|