Optimization of forecast of non-stationary objects based on fuzzy model adapters at external information influence
Dilshat Djumanov, Sunatillo Makhmudovich Kholmonov · Chemical Technology, Control and Management · 2020
The tasks of taking into account the sensitivity of fuzzy modeling based on the mechanisms for determining the range of elements of randomly chosen time series and the correction of the parameters of the functions of the accessories of linguistic variables, the use of the data property and the specific features of objects, the database and the knowledge base are solved. Computational schemes of dynamic identification based on polynomial models, nonlinear filters with fuzzy variable adapters are constructed. The effectiveness of generalized algorithms of fuzzy identification of randomly time series (RTS) is proved by comparison with the values of the characteristics of modal examples.