Self-spawning neuro-fuzzy system for rule extraction |
| |
Authors: | Zhi-Qiang Liu Tao Guan Ya-Jun Zhang |
| |
Affiliation: | (1) School of Creative Media, City University of Hong Kong, Hong Kong, Hong Kong SAR;(2) Department of Computer Science and Software Engineering, The University of Melbourne, Melbourne, VIC, 3010, Australia |
| |
Abstract: | In this paper we propose self-spawning neuro-fuzzy system (SSNFS), a new neuro-fuzzy system to derive fuzzy rules from data. The SSNFS model is based on a generic definition of incremental perceptron and a new learning algorithm that is capable of both structural (rule) learning and parametric learning. It constructs the fuzzy system by detecting a suitable number of rule patches and their positions and shapes in the input space. Initially the rule base consists of one single fuzzy rule; during the iterative learning process the rule base expands according to a supervised spawning validity measure. The rule induction process terminates when a given stop criterion is satisfied. SSNFS is very general since it does not require the prior knowledge about the input space and can be used in any application based on the scatter-partitioning fuzzy system. To demonstrate the effectiveness and applicability of our algorithm, we present a synthetic example and real-world modelling problems. |
| |
Keywords: | Incremental clustering Gaussian membership function Rule extraction Self-spawning competitive learning (SSCL) Self-spawning neuro-fuzzy system (SSNFS) |
本文献已被 SpringerLink 等数据库收录! |
|