Creating comprehensible regression models

Autor:	Johannes Himmelbauer, Mario Drobics
Rok vydání:	2006
Předmět:	Fuzzy classification Neuro-fuzzy business.industry Fuzzy set Machine learning computer.software_genre Defuzzification Fuzzy logic Theoretical Computer Science Fuzzy number Fuzzy set operations Geometry and Topology Artificial intelligence business computer Software Membership function Mathematics
Zdroj:	Soft Computing. 11:421-438
ISSN:	1433-7479 1432-7643
Popis:	In this paper we will present a novel approach to data-driven fuzzy modeling which aims to create highly accurate but also easily comprehensible models. This is achieved by a three-stage approach which separates the definition of the underlying fuzzy sets, the learning of the initial fuzzy model, and finally a local or global optimization of the resulting model. The benefit of this approach is that it allows to use a language comprising of comprehensible fuzzy predicates and to incorporate expert knowledge by defining problem specific fuzzy predicates. Furthermore, we achieve highly accurate results by applying a regularized optimization technique.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::dd29e66b9ffb07d1022927013963ab86 https://doi.org/10.1007/s00500-006-0107-1 Zobrazit plný text záznamu Full text from SpringerLink