Commutativity as prior knowledge in fuzzy modeling

被引：0

作者：

Carmona, P

Castro, JL

Zurita, JM

机构：

[1] Univ Extremadura, Dept Informat, Badajoz 06017, Spain

[2] Univ Granada, Dept Ciencias Computac & IA, ETSI Informat, E-18071 Granada, Spain

来源：

FUZZY SETS AND SYSTEMS - IFSA 2003, PROCEEDINGS | 2003年 / 2715卷

关键词：

fuzzy modeling; prior knowledge; commutativity;

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

This paper faces with the integration of mathematical properties satisfied by the system as prior knowledge in fuzzy modeling (FM), focusing on the commutativity as a starting point. The underlying idea is to reward the rules in each input fuzzy region that provide good commutativity degrees respecting its complementary -commutatively related- input fuzzy region. With this aim, the similarity between the outputs in both regions will be obtained. The experimental results show the accuracy improvement gained by the proposed method.

引用

页码：620 / 627

页数：8

共 12 条

[1]

ABUMOSTAFA Y, 1995, NEURAL NETWORKS CAPI, P221

[2] HINTS [J].