自组织神经网络在模糊聚类中的应用研究

聚类是按照事物的某些属性，把事物分类，使类间的相似性尽量小，类内的相似性尽量大。将事物通过适当聚类，才能便于研究事物的内部规律，但客观世界中存在着大量界线不分明的问题，研究模糊聚类的方法正是为了解决这类问题。在对常规模糊聚类方法分析的基础上，提出了一种将自组织竞争神经网络技术运用于模糊聚类的一种方法，并以100种动物分类为例，进行了模拟试验，仿真结果证明这种方法进行模糊聚类的思想正确，方法可行，效果较好。

Application of Self- Organizing Neural Networks in Fuzzy Clustering

Clustering is to classify thing in term of some property of thing, it can make the comparability least among kinds and comparability most in kinds. It is convenient for us to research the rule inside the thing to classify thing properly. Because there are many problems that borderline is not clear, the way of fuzzy clustering we research just about these problems. In this article, a new method that brings self- organizing competitive neural networks into the fields of fuzzy clustering is proposed on the basis of analysis the way of common fuzzy clustering,as a result,an example that realizes the classify of 100 animals is presented. The result of experiment shows this idea of fuzzy clustering is right, this way is capable and the result is better.