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