一种基于椭圆区域的进化式模糊分类系统

A Evolving Fuzzy Classifier System Based on Ellipsoidal Regions

介绍了一种进化式模糊分类系统。首先,介绍系统的基本特征及结构框架。然后,介绍了一种动态聚类算法,并运用动态聚类算法对输入的训练模式进行动态聚类,每一簇创建一条模糊规则。规则所对应的区域为类椭圆形区域。规则调整的策略是连续改变模糊分类规则的一个参数,使得分类系统对训练模式识别率不能再提高,对不能达到要求的调整,采用遗传算法进行调整。分析了规则调整的方法,给出了调整算法,也介绍了规则的插入和聚合策略,用两个典型的数据集来评测研究的系统,研究的分类系统在识别率与多层神经网络分类器相当,但训练时间远少于多层神经网络分类器的训练时间。

This paper introduces an evolving fuzzy classifier system. At first, the basic characteristics and frame of this system are introduced . Then , the dynamic clustering arithmetic which can dynamically cluster the input training patterns is presented. For every cluster, a fuzzy rule with an ellipsoidal region around a cluster center is defined. The strategy of tuning fuzzy rules is that the slopes of the membership functions are tuned successively until there is no improvement in the recognition rate of the training patterns . If this tuning can not satisfy the request , Genetic Algorithms will be used. In this paper , the tuning method and arithmetic, the policy of inserting rules and aggregating rules are discussed. This method has been evaluated by two typical data sets. The recognition rates of our classifier are comparable to the maximum recognition rates of the multilayered neural network classifier, and its training time is much shorter.