Premium
H ∞ identification and model structure selection
Author(s) -
Giarrè L.,
Milanese M.
Publication year - 1996
Publication title -
international journal of robust and nonlinear control
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.361
H-Index - 106
eISSN - 1099-1239
pISSN - 1049-8923
DOI - 10.1002/(sici)1099-1239(199605)6:4<367::aid-rnc238>3.0.co;2-j
Subject(s) - parametric statistics , identification (biology) , parametric model , set (abstract data type) , model selection , selection (genetic algorithm) , class (philosophy) , measure (data warehouse) , information criteria , radius , computer science , mathematics , goodness of fit , mathematical optimization , algorithm , data mining , artificial intelligence , machine learning , statistics , botany , computer security , biology , programming language
The advantages of a mixed parametric and non‐parametric approach, over the non‐parametric one, have been investigated in H ∞ set membership identification setting. The problem of evaluating the minimal worst case identification error, called radius of information, is solved. In particular, it is shown that the radius of information represents a measure of the ‘predictive ability’ of the considered class of models, and it is used to compare the ‘goodness’ of different classes of models and to choose the model order. Some numerical examples, showing the interest of the proposed test, are reported.