基于粗糙集理论的模糊聚类算法研究

Research on Fuzzy Clustering Algorithm Based on Rough Set Theory

粗糙集理论是一种新型的处理含糊不确定知识的数学工具,善于分析隐藏在数据中的事实而不需要关于数据的任何附加知识,粗集理论不仅为信息科学和认知科学提供了新的科学逻辑和研究方法,而且为智能信息处理提供了有效的处理技术。聚类是作为数据挖掘系统中的一个模块,既可以作为一个单独的工具以发现数据库中数据分布的深层信息,也可以作为其他数据挖掘分析算法的一个预处理步骤。模糊聚类算法忽略了聚类边界不确定的问题和复杂数据问题从而导致聚类效果不理想。本文提出了将粗糙集和模糊聚类算法相结合,利用粗糙集中上近似集和下近似集的概念得到相似性度量来改进模糊聚类算法。实验证明,改进的算法能够得到更好的聚类效果。

Rough set theory is a new mathematical tool for dealing with vague and uncertain knowledge.It is good at analyzing the facts hidden in the data without any additional knowledge about the data.Rough set theory not only provides new scientific logic and research methods for information science and cognitive science,but also provides effective processing technology for intelligent information processing.Clustering is a module in the data mining system.It can be used as a separate tool to discover the deep information of data distribution in the database,or as a pre-processing step of other data mining analysis algorithms.The fuzzy clustering algorithm ignores the problem of cluster boundary uncertainty and complex data problems,which leads to the unsatisfactory clustering effect.This paper proposes to combine the rough set and the fuzzy clustering algorithm,and use the concept of the approximate set and the lower approximation set on the rough set to obtain the similarity measure to improve the fuzzy clustering algorithm.Experiments show that the improved algorithm can get better clustering effect.