摘要
提出了一种改进的模糊分类系统的建模方法,采用模糊C均值聚类完成初始模糊分类系统的设计。提出改进的模糊规则置信度计算方法,对隶属函数和模糊规则相似度进行检测,剔除模糊规则中的冗余信息,利用遗传算法进行模糊分类系统的优化,提高系统的精确性和解释性。仿真结果证明了方法的有效性,对纤维图像的分类结果显示,该方法能获得与手工分类基本一致的分类结果。
An improved modeling method of fuzzy classifing system is proposed in the paper. The fuzzy C means clustering approach is adopted to design the initial fuzzy classifing system and the improved calculating approach of certainty degree is proposed. In order to remove the redundant information in the fuzzy rules, the similarities of the membership functions and the fuzzy rules are tested, and the fuzzy classifing system is optimized by using genetic algorithm to improve the accuracy and the interpretability of the system. The simulation result shows the validity of the proposed method, and the classifing results of fiber image show that the results obtained are similar to the manual classifiing results.
出处
《华东理工大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2008年第3期417-421,共5页
Journal of East China University of Science and Technology
基金
华东理工大学青年骨干教师基金项目(2007-03)
关键词
模糊分类系统
隶属函数
模糊规则
遗传算法
纤维图像
fuzzy classification system
membership function
fuzzy rules
genetic algorithm
fiber image
