F-transformācija un tās modifikācijas
Loading...
Date
Authors
Advisor
Journal Title
Journal ISSN
Volume Title
Publisher
Latvijas Universitāte
Language
N/A
Abstract
Darbs tiek veltīts F-transformācijai (F no angļu valodas „fuzzy” – nestrikts). Transformācija ir cieši saistīta ar nestriktiem sadalījumiem, kuru konstruēšanai darbā tiek izmantotas trīsstūrveida, trigonometriska veida un paraboliska veida harakteristiskās funkcijas. Ar tiešo F-transformāciju funkcija tiek pārvērsta par n-dimensionālu vektoru, kura komponentes apraksta transformējamās funkcijas vērtības atbilstoši intervāla [a,b] nestriktajam sadalījumam. Savukārt, ar inverso F-transformāciju n-dimensionāls vektors tiek pārvērsts atpakaļ par sākotnējās funkcijas aproksimāciju. Gan tiešā, gan inversā H-transformācija tiek aprakstīta kā F-transformācijas modifikācija. Abu transformāciju gadījumā tiek iegūti kļūdu novērtējumi. Rezultāti ilustrēti ar testa piemēriem.
This paper is devoted to F-transform (fuzzy transform). The core idea of transform is inwrought with an interval [a,b] fuzzy partitioning into fuzzy subsets, determined by their membership functions. In this work we consider triangular, sinusoidal and quadratic membership functions. The idea of F-transform is transformation from a function space to a finite dimensional vector space. The inverse F-transform is transformation back to the function space. We claim that for a sufficient representation of a function defined on [a,b], we may consider its average values over fuzzy subsets from the partition. The modification of F-transform (named by H-transform) is introduced in this work. Error bounds for both transforms are obtained and analysed with test examples.
This paper is devoted to F-transform (fuzzy transform). The core idea of transform is inwrought with an interval [a,b] fuzzy partitioning into fuzzy subsets, determined by their membership functions. In this work we consider triangular, sinusoidal and quadratic membership functions. The idea of F-transform is transformation from a function space to a finite dimensional vector space. The inverse F-transform is transformation back to the function space. We claim that for a sufficient representation of a function defined on [a,b], we may consider its average values over fuzzy subsets from the partition. The modification of F-transform (named by H-transform) is introduced in this work. Error bounds for both transforms are obtained and analysed with test examples.